JP2015104981A - Vehicular acceleration suppression apparatus and vehicular acceleration suppression method - Google Patents

Abstract

PROBLEM TO BE SOLVED: To provide a drive support appropriate for an erroneous operation of an accelerator of a driver.SOLUTION: In a case where an edge candidate being a candidate of an edge portion constituting a parking zone and a numeral zone (NA) including a symbol string containing a numeral are extracted from a pickup image produced by picking up a zone including a road surface surrounding a vehicle (V), with a distance between the extracted edge candidate and numeral zone (NA) within a pre-set first distance, a parking zone constituted by the edge candidate within the first distance and the numeral zone (NA) is detected. Then, upon detecting the parking zone, acceleration is reduced, the acceleration generated for the vehicle (V) in accordance with an operation amount of an acceleration pedal (32) of a driver.

Description

  The present invention relates to a technique for assisting driving when a vehicle is parked.

  As a device for controlling the speed of a vehicle, for example, there is a safety device described in Patent Document 1. In this safety device, it is detected from the map data of the navigation device and the current position information that the vehicle is off the road, there is an accelerator operation in a direction to increase the vehicle traveling speed, and the vehicle traveling speed is predetermined. When it is determined that the value is larger than the value of the throttle, the throttle is controlled in the deceleration direction regardless of the operation of the accelerator.

JP 2003-137001 A

The prior art disclosed in Patent Document 1 aims to prevent acceleration of the vehicle that is not intended by the driver even if the accelerator operation is erroneously performed. At this time, it becomes a problem to determine whether or not the accelerator operation is an erroneous operation. And in the above-mentioned Patent Document 1, it is assumed that the accelerator depressing operation when the host vehicle is at a position deviating from the road based on the map information and the traveling speed of a predetermined value or more is detected may be an accelerator misoperation, The above conditions are operating conditions for throttle suppression.
However, under the above-described operating conditions, the throttle control is activated depending on the vehicle speed only by getting off the road and entering the parking lot, and the drivability in the parking lot is reduced.
The present invention has been made paying attention to the above points, and an object of the present invention is to suppress acceleration of the vehicle that is not intended by the driver when the host vehicle is parked in a parking area such as a parking frame.

  In order to solve the above-described problem, according to one aspect of the present invention, an end candidate that is a candidate for an end that forms a parking area, and a symbol that includes a number from a captured image obtained by capturing an area including a road surface around the host vehicle Extract the numeric area containing the column. In addition, when the distance between the extracted edge candidate and the extracted number area is within the first distance range set in advance, it is composed of the edge candidate and the number area that are within the first distance range. Detect the parking area. When the parking area is detected, the acceleration generated in the host vehicle is reduced according to the acceleration operation amount of the driver's acceleration operation element.

One aspect of the present invention detects a parking area composed of an end candidate and a numeric area whose distance from a numeric area located on the road surface is within a preset first distance range. When the parking area is detected, control for reducing acceleration generated in the host vehicle according to the acceleration operation amount (for example, the operation amount of the accelerator pedal) of the driver's acceleration operation element (in the following description, “acceleration suppression control”). May be written).
As a result, when the host vehicle is traveling with respect to the parking area, if an accelerator operation error occurs, the acceleration of the host vehicle is reduced, so the driver when the host vehicle parks in the parking area. It is possible to suppress the occurrence of unintended acceleration.

It is a conceptual diagram which shows the structure of a vehicle provided with the acceleration suppression apparatus for vehicles. It is a block diagram which shows schematic structure of the acceleration suppression apparatus for vehicles. It is a block diagram which shows the structure of the acceleration suppression control content calculating part. 4 is a block diagram illustrating an example of a functional configuration of a parking area certainty setting unit 36. FIG. It is a figure which shows the pattern of the parking area which the parking area reliability setting part 36 makes the setting object of parking area reliability. It is a figure which shows the pattern of the parking area which the parking area reliability setting part 36 makes the setting object of parking area reliability. It is a flowchart which shows the process which an acceleration suppression operation condition judgment part judges whether an acceleration suppression operation condition is materialized. It is a figure explaining the distance of the own vehicle, a parking area, and the own vehicle and a parking area. It is a flowchart which shows the process which a parking area reliability calculation part sets a parking area reliability. It is a flowchart which shows an example of the process sequence of parking area | region adaptation condition determination processing. (A) And (b) is a schematic diagram which illustrates typically the extraction method of the edge part candidate by edge detection. (A)-(c) is a schematic diagram which shows the example in the case where the space | interval of two edge part candidates is within the preset space | interval range, and the case where it is outside a space | interval range. (A)-(d) is a schematic diagram which shows the example in the case where it does not match | combine with the case where the shape of two edge part candidates is adapted to the combination of the shape defined beforehand. (A)-(f) is a schematic diagram which shows the example in the case where there exists a case where there is no shift | offset | difference of the direction of an edge part candidate pair. (A) And (b) is a schematic diagram which illustrates typically the extraction method of the parking area | region line candidate by edge detection. It is a figure which shows the content of the process which the parking area | region detection part 36d performs. It is a flowchart which shows an example of the process sequence of a 1st number determination process. (A)-(h) is a figure which shows an example of the combination of the shape of a parking area | region, and the position of identification information. It is a figure which shows an example of the distance relationship between the edge part which comprises a parking area | region, and a number area. (A) is a figure which shows an example of the threshold value map for edge detection, (b) is a figure which shows an example of the matching threshold value map for pattern matching. (A)-(c) is a figure which shows an example of the number used on the public road. (A)-(e) is a figure which shows an example of the dimension and distance referred when performing exclusion determination of a number area | region, (f) is a figure which shows an example of the peripheral area | region of a number area | region. It is a flowchart which shows an example of the process sequence of a 2nd number determination process. It is a figure which shows an example of the distance relationship between the parking area line and number area which comprise a parking area. (A) is a figure which shows an example of the determination distance when the extracted parking area line candidate is one, (b) is a figure which shows an example of the determination distance when an area line candidate pair is extracted. (C) is a figure which shows an example of the shift | offset | difference from the reference | standard direction of a number area | region and a parking area | region line candidate. (A) is a figure which shows an example of the horizontal width | variety of a number area | region, and the width | variety between area | region line candidate pairs, (b) is a figure which shows an example of an area | region line candidate pair and the peripheral area | region of a number area | region. It is a figure which shows the content of the process which the parking area reliability calculation part 36e performs. It is a flowchart which shows an example of the process sequence which a parking area approach reliability setting part sets a parking area approach reliability. It is a figure which shows the content of the process which detects the deviation | shift amount of the prediction locus | trajectory of the own vehicle, and a parking area. It is a figure which shows an example of a comprehensive reliability setting map. It is a figure which shows an example of an acceleration suppression condition calculation map. It is a flowchart which shows an example of the process sequence of the process which an acceleration suppression command value calculating part performs. It is a flowchart which shows an example of the process sequence of the process which a target throttle opening calculating part performs. (A) is a figure which shows an example when only the right front end part of the parking area comprised only by an L-shaped edge part is extracted, (b) is comprised from two parking area lines. It is a figure which shows an example when only the parking area line of the right side of a parking area is extracted, (c) is a figure which shows an example of the extracted number area and a non-number area. It is a figure which shows the modification in the case of performing distance determination with a number area | region with respect to the parking area of the structure which does not contain an identification number in a frame. It is a figure which shows the modification of an acceleration suppression condition calculation map.

Embodiments of the present invention will be described below with reference to the drawings.
(Constitution)
First, the configuration of a vehicle including the vehicle acceleration suppression device of the present embodiment will be described with reference to FIG.
As shown in FIG. 1, the vehicle V includes wheels W (a right front wheel WFR, a left front wheel WFL, a right rear wheel WRR, a left rear wheel WRL), a brake device 2, a fluid pressure circuit 4, and a brake controller 6. . In addition to this, the vehicle V includes an engine 8 and an engine controller 12.
The brake device 2 is formed using, for example, a wheel cylinder and provided on each wheel W. The brake device 2 is not limited to a device that applies a braking force with fluid pressure, and may be formed using an electric brake device or the like.

The fluid pressure circuit 4 is a circuit including piping connected to each brake device 2.
The brake controller 6 responds to the braking force command value generated by each brake device 2 via the fluid pressure circuit 4 based on the braking force command value received from the travel controller 10 that is the host controller. To control the value. That is, the brake controller 6 forms a deceleration control device. In addition, the description regarding the traveling control controller 10 is mentioned later.
Therefore, the brake device 2, the fluid pressure circuit 4, and the brake controller 6 form a braking device that generates a braking force.

The engine 8 forms a drive source for the vehicle V.
The engine controller 12 controls the torque (driving force) generated by the engine 8 based on the target throttle opening signal (acceleration command value) received from the travel controller 10. That is, the engine controller 12 forms an acceleration control device. A description regarding the target throttle opening signal will be given later.
Therefore, the engine 8 and the engine controller 12 form a driving device that generates driving force.
The drive source of the vehicle V is not limited to the engine 8, and may be formed using an electric motor. Further, the drive source of the vehicle V may be formed by combining the engine 8 and an electric motor.

Next, a schematic configuration of the vehicle acceleration suppression device 1 will be described with reference to FIG.
As shown in FIGS. 1 and 2, the vehicle acceleration suppression device 1 includes an ambient environment recognition sensor 14, a wheel speed sensor 16, a steering angle sensor 18, a shift position sensor 20, and a brake operation detection sensor 22. And an accelerator operation detection sensor 24. In addition, the vehicle acceleration suppression device 1 includes a navigation device 26 and a travel controller 10.
The ambient environment recognition sensor 14 captures an image around the vehicle V, and based on each captured image, an information signal including individual images corresponding to a plurality of imaging directions (in the following description, “individual image signal”). May be listed). Then, the generated individual image signal is output to the travel controller 10.

In the present embodiment, as an example, a case where the surrounding environment recognition sensor 14 is formed using a front camera 14F, a right side camera 14SR, a left side camera 14SL, and a rear camera 14R will be described. Here, the front camera 14 </ b> F is a camera that images the front of the vehicle V in the vehicle front-rear direction, and the right-side camera 14 </ b> SR is a camera that images the right side of the vehicle V. Further, the left side camera 14SL is a camera that images the left side of the vehicle V, and the rear camera 14R is a camera that images the rear of the vehicle V in the vehicle front-rear direction.
In the present embodiment, the ambient environment recognition sensor 14 captures a distance range of the maximum shooting range (for example, 100 [m]) of each camera at an angle of view where a road surface around the vehicle V enters, for example.

The wheel speed sensor 16 is formed using, for example, a pulse generator such as a rotary encoder that measures wheel speed pulses.
Further, the wheel speed sensor 16 detects the rotational speed of each wheel W, and an information signal including the detected rotational speed (which may be referred to as “wheel speed signal” in the following description) is used as a travel controller. 10 is output.
For example, the steering angle sensor 18 is provided in a steering column (not shown) that rotatably supports the steering wheel 28.
The steering angle sensor 18 detects a current steering angle that is a current rotation angle (steering operation amount) of the steering wheel 28 that is a steering operator, and an information signal including the detected current steering angle (in the following description). , May be described as “current steering angle signal”) to the travel controller 10. In addition, you may detect the information signal containing the steering angle of a steered wheel as information which shows a steering angle.

Further, the steering operator is not limited to the steering wheel 28 that is rotated by the driver, and may be, for example, a lever that is operated by the driver to tilt by hand. In this case, the lever tilt angle from the neutral position is output as an information signal corresponding to the current steering angle signal.
The shift position sensor 20 detects the current position of a member that changes the shift position (for example, “P”, “D”, “R”, etc.) of the vehicle V, such as a shift knob or a shift lever. Then, an information signal including the detected current position (which may be described as a “shift position signal” in the following description) is output to the travel controller 10.
The brake operation detection sensor 22 detects the opening degree of the brake pedal 30 that is a braking force instruction operator. Then, an information signal including the detected opening degree of the brake pedal 30 (which may be described as a “brake opening degree signal” in the following description) is output to the travel controller 10.

Here, the braking force instruction operator is configured to be operable by the driver of the vehicle V and to instruct the braking force of the vehicle V by changing the opening. Note that the braking force instruction operator is not limited to the brake pedal 30 that the driver steps on with his / her foot, and may be, for example, a lever that is manually operated by the driver.
The accelerator operation detection sensor 24 detects the opening degree of the accelerator pedal 32 that is a driving force instruction operator. Then, an information signal including the detected opening of the accelerator pedal 32 (in the following description, it may be described as “accelerator opening signal”) is output to the travel controller 10.

Here, the driving force indicating operator is configured to be operable by the driver of the vehicle V and to instruct the driving force of the vehicle V by a change in the opening degree. Note that the driving force instruction operator is not limited to the accelerator pedal 32 on which the driver performs the stepping operation with his / her foot, and may be, for example, a lever operated by the driver with his hand.
The navigation device 26 includes a GPS (Global Positioning System) receiver, a map database, and an information presentation device having a display monitor and the like, and performs route search, route guidance, and the like.

The navigation device 26 also obtains road information such as the type and width of the road on which the vehicle V travels based on the current position of the vehicle V acquired using the GPS receiver and the road information stored in the map database. It is possible to obtain.
In addition, the navigation device 26 sends an information signal (which may be referred to as “own vehicle position signal” in the following description) including the current position of the vehicle V acquired using the GPS receiver to the travel controller 10. Output. In addition to this, the navigation device 26 outputs an information signal including the type of road on which the vehicle V travels, the road width, and the like (in the following description, it may be described as “traveling road information signal”) to the travel controller 10. Output to.

The information presenting device outputs an alarm or other presenting by voice or image in accordance with a control signal from the travel controller 10. In addition, the information presentation device includes, for example, a speaker that provides information to the driver by a buzzer sound or voice, and a display unit that provides information by displaying an image or text. Further, the display unit may divert the display monitor of the navigation device 26, for example.
The travel control controller 10 is an electronic control unit that includes a CPU and CPU peripheral components such as a ROM and a RAM.
The travel controller 10 also includes a parking driving support unit that performs driving support processing for parking.

  As shown in FIG. 2, the parking driving support unit among the processes of the travel controller 10 includes, as shown in FIG. 2, an ambient environment recognition information calculation unit 10A, a host vehicle speed calculation unit 10B, and a steering angle calculation unit 10C. A steering angular velocity calculation unit 10D. In addition, the parking driving support unit includes a shift position calculation unit 10E, a brake pedal operation information calculation unit 10F, an accelerator operation amount calculation unit 10G, an accelerator operation speed calculation unit 10H, and an acceleration suppression control as functional components. A content calculation unit 10I. Furthermore, the parking driving support unit includes an acceleration suppression command value calculation unit 10J and a target throttle opening calculation unit 10K as functional components. These functional components are composed of one or more programs.

The ambient environment recognition information calculation unit 10A forms an image (a bird's-eye view image) around the vehicle V as viewed from above the vehicle V based on the individual image signal received from the ambient environment recognition sensor 14. Then, an acceleration suppression control content calculation unit 10I includes an information signal including the formed overhead image (in the following description, may be referred to as “overhead image signal”) and an individual image signal corresponding to the overhead image signal. Output to.
Here, the bird's-eye view image is formed by, for example, synthesizing images captured by the respective cameras (front camera 14F, right side camera 14SR, left side camera 14SL, and rear camera 14R). In addition, in the bird's-eye view image, for example, an end portion constituting a parking area such as a parking frame, which is marked on the road surface (in the following description, may be described as “parking area end”), a parking area It includes a marking image such as a line (which may be described as “parking area line” in the following description).

In the present embodiment, the parking area is an area where the vehicle is individually parked, and is an area partitioned by a parking area end or a parking area line marked on a road surface in a predetermined site such as a parking lot. is there. Some parking areas have identification information (such as an identification number) in the vicinity (for example, marked on the road surface) in order to identify each parking area.
Further, in the present embodiment, the parking area end corresponds to, for example, a marking pattern having a predetermined shape that is arranged at the end of the parking area line, the four corners of the parking area, or the like and divides the parking area.
In the present embodiment, the parking area line includes, for example, a frame-shaped line (for example, a rectangular shape) that partitions the parking area, two parallel straight lines that partition the parking area, and two parallel lines that partition the parking area. For example, a U-shaped line composed of a straight line and one straight line orthogonal to these two straight lines is applicable.

The own vehicle vehicle speed calculation unit 10 </ b> B calculates the speed (vehicle speed) of the vehicle V from the rotation speed of the wheel W based on the wheel speed signal received from the wheel speed sensor 16. Then, an information signal including the calculated speed (in the following description, may be described as “vehicle speed calculation value signal”) is output to the acceleration suppression control content calculation unit 10I.
The steering angle calculation unit 10C calculates the operation amount (rotation angle) from the neutral position of the steering wheel 28 from the current rotation angle of the steering wheel 28 based on the current steering angle signal received from the steering angle sensor 18. . Then, an information signal including the calculated operation amount from the neutral position (in the following description, may be described as “steering angle signal”) is output to the acceleration suppression control content calculation unit 10I.

The steering angular velocity calculation unit 10D calculates the steering angular velocity of the steering wheel 28 by differentiating the current steering angle included in the current steering angle signal received from the steering angle sensor 18. Then, an information signal including the calculated steering angular velocity (may be described as “steering angular velocity signal” in the following description) is output to the acceleration suppression control content calculation unit 10I.
The shift position calculation unit 10E determines the current shift position based on the shift position signal received from the shift position sensor 20. Then, an information signal including the calculated current shift position (in the following description, may be described as “current shift position signal”) is output to the acceleration suppression control content calculation unit 10I.

The brake pedal operation information calculation unit 10F calculates the depression amount of the brake pedal 30 based on a state where the depression amount is “0” based on the brake opening signal received from the brake operation detection sensor 22. Then, an information signal including the calculated depression amount of the brake pedal 30 (in the following description, may be described as a “braking side depression amount signal”) is output to the acceleration suppression control content calculation unit 10I.
The accelerator operation amount calculation unit 10G calculates the depression amount of the accelerator pedal 32 with reference to the state where the depression amount is “0” based on the accelerator opening signal received from the accelerator operation detection sensor 24. Then, an information signal including the calculated depression amount of the accelerator pedal 32 (in the following description, may be described as a “driving-side depression amount signal”), an acceleration suppression control content calculation unit 10I, and an acceleration suppression command value calculation To the unit 10J and the target throttle opening calculation unit 10K.

The accelerator operation speed calculation unit 10H calculates the operation speed of the accelerator pedal 32 by differentiating the opening of the accelerator pedal 32 included in the accelerator opening signal received from the accelerator operation detection sensor 24. Then, an information signal including the calculated operation speed of the accelerator pedal 32 (in the following description, may be described as “accelerator operation speed signal”) is output to the acceleration suppression command value calculation unit 10J.
The acceleration suppression control content calculation unit 10I includes the above-described various information signals (overhead image signal, individual image signal, vehicle speed calculation value signal, steering angle signal, steering angular velocity signal, current shift position signal, braking side depression amount signal, driving side Input of a depression amount signal, a vehicle position signal, and a traveling road information signal. And based on the various information signals which received the input, the acceleration suppression operation condition judgment result mentioned later, acceleration suppression control start timing, and acceleration suppression control amount are calculated. Furthermore, an information signal including these calculated parameters is output to the acceleration suppression command value calculation unit 10J.
The detailed configuration of the acceleration suppression control content calculation unit 10I and the processing performed by the acceleration suppression control content calculation unit 10I will be described later.

The acceleration suppression command value calculation unit 10J inputs the drive side depression amount signal and the accelerator operation speed signal described above, and inputs an acceleration suppression operation condition determination result signal, an acceleration suppression control start timing signal, and an acceleration suppression control amount signal described later. receive. Then, an acceleration suppression command value, which is a command value for reducing the acceleration generated in the vehicle V, is calculated according to the depression amount (acceleration operation amount) of the accelerator pedal 32. Further, an information signal including the calculated acceleration suppression command value (in the following description, may be described as “acceleration suppression command value signal”) is output to the target throttle opening calculation unit 10K.
Further, the acceleration suppression command value calculation unit 10J calculates a normal acceleration command value, which is a command value used in normal acceleration control, according to the content of the received acceleration suppression operation condition determination result signal. Furthermore, an information signal including the calculated normal acceleration command value (in the following description, may be described as “normal acceleration command value signal”) is output to the target throttle opening calculation unit 10K.
The processing performed by the acceleration suppression command value calculation unit 10J will be described later.

The target throttle opening calculation unit 10K receives an input of a drive side depression amount signal and an acceleration suppression command value signal. Based on the depression amount of the accelerator pedal 32 and the acceleration suppression command value, a target throttle opening that is a throttle opening corresponding to the depression amount of the accelerator pedal 32 or the acceleration suppression command value is calculated. Further, an information signal including the calculated target throttle opening (in the following description, may be described as “target throttle opening signal”) is output to the engine controller 12.
Further, when the acceleration suppression command value includes an acceleration suppression control start timing command value described later, the target throttle opening calculation unit 10K sends the target throttle opening signal to the engine controller 12 based on the acceleration suppression control start timing described later. Output.
The processing performed by the target throttle opening calculation unit 10K will be described later.

(Configuration of acceleration suppression control content calculation unit 10I)
Next, a detailed configuration of the acceleration suppression control content calculation unit 10I will be described using FIGS. 3 to 6 with reference to FIGS.
As shown in FIG. 3, the acceleration suppression control content calculation unit 10I includes an acceleration suppression operation condition determination unit 34, a parking area certainty level setting unit 36, a parking area entry certainty level setting unit 38, and an overall reliability level setting unit. 40. In addition, the acceleration suppression control content calculation unit 10I includes an acceleration suppression control start timing calculation unit 42 and an acceleration suppression control amount calculation unit 44.

The acceleration suppression operation condition determination unit 34 determines whether or not a condition for operating acceleration suppression control is satisfied, and describes an information signal including the determination result (in the following description, “acceleration suppression operation condition determination result signal”). Is output to the acceleration suppression command value calculation unit 10J. Here, the acceleration suppression control is control for setting an acceleration command value for accelerating the vehicle V in accordance with the amount of depression of the accelerator pedal 32 to a value that reduces acceleration more than usual. In this embodiment, since acceleration increases as the acceleration command value increases, the acceleration suppression control is control that reduces the acceleration command value for accelerating the vehicle V in accordance with the depression amount of the accelerator pedal 32 than usual.
Moreover, the process in which the acceleration suppression operation condition determination part 34 determines whether the conditions for operating the acceleration suppression control are satisfied will be described later.

The parking area certainty setting unit 36 sets a parking area certainty that is a certainty that a parking area exists in the traveling direction of the vehicle V. Then, an information signal including the set parking area certainty factor (may be described as a “parking area certainty signal” in the following description) is output to the total certainty factor setting unit 40.
Here, the parking area certainty setting unit 36 refers to various information included in the overhead image signal, the individual image signal, the vehicle speed calculation value signal, the current shift position signal, the own vehicle position signal, and the traveling road information signal, Set confidence.
The process in which the parking area certainty setting unit 36 sets the parking area certainty will be described later.

The parking area approach certainty setting unit 38 sets a parking area approach certainty that is a certainty that the vehicle V enters the parking area. Then, an information signal including the set parking area approach certainty factor (may be described as “parking area approach certainty signal” in the following description) is output to the total confidence setting unit 40.
Here, the parking area approach certainty setting unit 38 sets the parking area approach certainty with reference to various information included in the overhead image signal, the vehicle speed calculation value signal, the current shift position signal, and the steering angle signal.
In addition, the process which the parking area approach reliability setting part 38 sets a parking area approach reliability is mentioned later.

The comprehensive certainty setting unit 40 receives the input of the parking area certainty signal and the parking area approach certainty signal, and sets the overall certainty that is the certainty corresponding to the parking area certainty and the parking area approach certainty. Then, an information signal including the set total certainty factor (may be described as a “total confidence signal” in the following description) is output to the acceleration suppression control start timing calculation unit 42 and the acceleration suppression control amount calculation unit 44. To do.
In addition, the process which the comprehensive reliability setting part 40 sets a comprehensive reliability is mentioned later.

The acceleration suppression control start timing calculation unit 42 calculates an acceleration suppression control start timing that is a timing for starting the acceleration suppression control. Then, an information signal including the calculated acceleration suppression control start timing (may be described as “acceleration suppression control start timing signal” in the following description) is output to the acceleration suppression command value calculation unit 10J.
Here, the acceleration suppression control start timing calculation unit 42 refers to various information included in the comprehensive reliability signal, the braking side depression amount signal, the vehicle speed calculation value signal, the current shift position signal, and the steering angle signal, and starts the acceleration suppression control. Calculate timing.
The process in which the acceleration suppression control start timing calculation unit 42 calculates the acceleration suppression control start timing will be described later.

The acceleration suppression control amount calculation unit 44 calculates an acceleration suppression control amount that is a control amount for reducing the acceleration command value corresponding to the depression amount of the accelerator pedal 32. Then, an information signal including the calculated acceleration suppression control amount (in the following description, may be described as “acceleration suppression control amount signal”) is output to the acceleration suppression command value calculation unit 10J.
Here, the acceleration suppression control amount calculation unit 44 refers to various information included in the comprehensive certainty signal, the braking side depression amount signal, the vehicle speed calculation value signal, the current shift position signal, and the steering angle signal, and determines the acceleration suppression control amount. Calculate.
The process in which the acceleration suppression control amount calculation unit 44 calculates the acceleration suppression control amount will be described later.

(Configuration of parking area certainty setting unit 36)
Next, a detailed configuration of the parking area certainty factor setting unit 36 will be described with reference to FIGS. 1 to 3 and FIGS. 4 to 6.
As shown in FIG. 4, the parking area certainty level setting unit 36 includes an end candidate extracting unit 36a, a line extracting unit 36b, a numeric area extracting unit 36c, a parking area detecting unit 36d, and a parking area certainty factor calculating unit. 36e.
Here, as shown in FIG. 5 and FIG. 6, for example, there are a plurality of patterns in the parking area that the parking area certainty level setting unit 36 sets the certainty level.
That is, the parking area is divided into, for example, two parallel straight lines (vertical lines) shown in FIG. 5 (a), two parallel vertical lines as shown in FIG. There are one that is partitioned in a “U” shape by one horizontal line orthogonal to the vertical line, one that is partitioned by a rectangular frame line shown in FIG. 5C (so-called parking frame), and the like. In addition, there are various patterns as shown in FIGS. 5 (d) to 5 (q). Moreover, as shown in FIG.5 (n) and FIG.5 (q), for example, as shown in FIG.5 (n) and FIG.5 (q), there are some parking areas divided by the parking area line which is common with the adjacent parking area.

Further, the parking area is, for example, divided by only four L-shaped marking patterns (parking area end portions) arranged at the four corners shown in FIG. 6A, and the four corners shown in FIG. 6B. And the like, which are partitioned by only four U-shaped marking patterns (parking region end portions) arranged in the area.
The edge candidate extraction unit 36a may be described as an edge candidate that constitutes a parking area located on the road surface from the bird's-eye view image of the vehicle V in the vehicle traveling direction (in the following description, “end candidate”). ).

  Specifically, the end candidate extraction unit 36a ends an image area that meets the conditions as an end of the parking area line (for example, an end of a line constituting the parking area shown in FIG. 5) from the overhead image. Extracted as a copy candidate. In addition, from a bird's-eye view image, a predetermined shaped marking pattern (for example, an L-shaped or U-shaped marking pattern shown in FIG. 6) is arranged at the four corners of the road surface area constituting the parking area. Image regions that meet the above conditions are extracted as edge candidates. Then, the edge candidate extraction unit 36a uses the number area extraction unit 36c and the parking area detection as information on the extraction results of the edge candidates (in the following description, may be referred to as “edge candidate extraction result information”). To the unit 36d.

The line extraction unit 36b obtains a candidate for a line constituting a parking area located on the road surface from the overhead image in the vehicle traveling direction of the vehicle V (may be described as “parking area line candidate” in the following description). Extract.
Specifically, the line extraction unit 36b selects a line that matches a condition as a parking area line (for example, a line constituting the parking area shown in FIG. 5) from among the lines located on the road surface as parking area line candidates. Extract as And the line extraction part 36b is the number area extraction part 36c and the parking area | region detection part for the information of the extraction result of a parking area line candidate (it may be described as "area line candidate extraction result information" in the following description). To 36d.

Based on the input edge candidate extraction result information and region line candidate extraction result information, the number region extraction unit 36c generates, from the overhead image, an image region portion including a symbol string including numbers (in the following description, “number region”). Is extracted). And the number area extraction part 36c outputs the information of the extraction result of a number area (it may be described as "number area extraction result information" in the following description) to the parking area detection part 36d.
Here, the number area to be extracted is an image area portion including a symbol string made up of numbers as the identification information (identification number) of the parking area or a character string made up of numbers and symbols other than numbers. The symbol string also includes a symbol string of one digit (one row and one column). For example, the identification number of the parking area is represented by a number “0-9” or a combination of these numbers, “A1”, “A2”, “light 1”, “light 2”, and the like. There are things such as alphabets, numbers and kanji, which are represented by combinations of numbers and symbols other than numbers.

The parking area detection unit 36d performs an edge determination process based on the input edge candidate extraction result information. The edge determination process specifies a pair of edge candidates that matches a condition as a pair of edges constituting a preset parking area from the extracted edge candidates. And it becomes the process which detects the parking area | region comprised from this identified edge part pair (it may describe as an "edge candidate pair" in the following description).
The parking area detection unit 36d further performs an area line determination process based on the input area line candidate extraction result information. In the area line determination process, a pair of parking area line candidates that match a condition as a pair of parking area lines constituting a preset parking area is specified from the extracted parking area line candidates. And it becomes the process which detects the parking area comprised from this identified parking area line candidate pair (in the following description, it may describe as an "area line candidate pair").

The parking area detection unit 36d further performs a first numeral determination process based on the input end candidate extraction result information and the numeral area extraction result information. The first number determination process is performed in such a manner that an end portion that matches a condition as a set of an end portion and a number region that constitute a parking area is selected from the extracted end portion candidates and the extracted number region. This is a process of detecting a pair of a candidate and a number area as a parking area.
The parking area detection unit 36d further performs a second number determination process based on the input area line candidate extraction result information and number area extraction result information. The second number determination process conforms to a predetermined condition as a set of a parking area line and a number area constituting the parking area from among the extracted parking area line candidates and the extracted number area. This is a process of detecting a parking area line candidate and a number area as a parking area.

The parking area detection unit 36d is information of processing results of the edge determination process, the area line determination process, the first numeral determination process, and the second numeral determination process (in the following description, “parking area detection result information” is described) Is output to the parking area certainty calculation unit 36e.
The parking area certainty calculation unit 36e calculates the parking area certainty based on the parking area detection result information. In addition, the detail about the process which calculates parking area reliability is mentioned later.

(Processing performed by the acceleration suppression control content calculation unit 10I)
Next, processing performed by the acceleration suppression control content calculation unit 10I will be described with reference to FIGS. 1 to 6 and FIGS.
Processing performed by the acceleration suppression operation condition determination unit 34 Referring to FIGS. 1 to 6, the conditions for the acceleration suppression operation condition determination unit 34 to operate the acceleration suppression control (refer to the following description). The process of determining whether or not “acceleration suppression operation condition” may be established will be described.
That is, the acceleration suppression operation condition determination unit 34 repeatedly performs the process described below every preset sampling time (for example, 10 [msec]).
As shown in FIG. 7, when the acceleration suppression operation condition determination unit 34 starts processing (START), first, the process proceeds to step S100.

In step S100, the acceleration suppression operation condition determination unit 34 performs a process of acquiring the parking area certainty set by the parking area certainty setting unit 36 ("parking area certainty acquisition process" shown in the figure). Thereafter, the processing performed by the acceleration suppression operation condition determination unit 34 proceeds to step S102.
In step S102, the acceleration suppression operation condition determination unit 34 performs a process of determining the presence / absence of the parking area based on the parking area certainty factor acquired in step S100 ("parking presence / absence determination process" shown in the figure).
In this embodiment, the process which judges the presence or absence of a parking area is performed based on a parking area reliability. Specifically, if it is determined that the parking area certainty is a preset minimum value (level 0), for example, there is no parking area within a distance or area (area) set in advance with reference to the vehicle V (see FIG. "No" shown in the figure). In this case, the process performed by the acceleration suppression operation condition determination unit 34 proceeds to step S120.

On the other hand, if it is determined that the certainty of the parking area is a value other than the preset minimum value, the parking area is within a distance or area (area) set in advance with reference to the vehicle V (“Yes” shown in the figure). )). In this case, the process performed by the acceleration suppression operation condition determination unit 34 proceeds to step S104.
In step S104, the acceleration suppression operation condition determination unit 34 refers to the vehicle speed calculation value signal received from the host vehicle speed calculation unit 10B and acquires the vehicle speed of the vehicle V ("host vehicle speed shown in the figure"). Information acquisition process ”). Thereafter, the processing performed by the acceleration suppression operation condition determination unit 34 proceeds to step S106.

In step S106, the acceleration suppression operation condition determination unit 34 determines whether or not the condition that the vehicle speed of the vehicle V is less than a preset threshold vehicle speed is satisfied based on the vehicle speed acquired in step S104 ( “Self-vehicle speed condition determination process” shown in FIG.
In the present embodiment, a case where the threshold vehicle speed is set to 15 [km / h] will be described as an example. The threshold vehicle speed is not limited to 15 [km / h], and may be changed according to the specifications of the vehicle V such as the braking performance of the vehicle V, for example. Further, for example, the vehicle V may be changed according to traffic regulations or the like of the region (country or the like) where the vehicle V travels.

If it is determined in step S106 that the condition that the vehicle speed of the vehicle V is less than the threshold vehicle speed is satisfied ("Yes" shown in the figure), the processing performed by the acceleration suppression operation condition determination unit 34 proceeds to step S108. To do.
On the other hand, if it is determined in step S106 that the condition that the vehicle speed of the vehicle V is less than the threshold vehicle speed is not satisfied ("No" shown in the drawing), the processing performed by the acceleration suppression operation condition determination unit 34 is performed in step S120. Migrate to
In step S108, the acceleration suppression operation condition determination unit 34 refers to the brake-side depression amount signal received from the brake pedal operation information calculation unit 10F, and acquires information on the depression amount (operation amount) of the brake pedal 30. Processing ("brake pedal operation amount information acquisition processing" shown in the figure) is performed. Thereafter, the processing performed by the acceleration suppression operation condition determination unit 34 proceeds to step S110.

In step S110, the acceleration suppression operation condition determination unit 34 determines whether or not the brake pedal 30 is operated based on the depression amount of the brake pedal 30 acquired in step S108 ("brake pedal shown in the figure"). Operation determination process ").
If it is determined in step S110 that the brake pedal 30 is not operated ("No" shown in the figure), the processing performed by the acceleration suppression operation condition determination unit 34 proceeds to step S112.
On the other hand, when it is determined in step S110 that the brake pedal 30 is operated (“Yes” shown in the drawing), the processing performed by the acceleration suppression operation condition determination unit 34 proceeds to step S120.

In step S112, the acceleration suppression operation condition determination unit 34 refers to the driving side depression amount signal received from the accelerator operation amount calculation unit 10G and acquires information on the depression amount (operation amount) of the accelerator pedal 32. ("Accelerator pedal operation amount information acquisition process" shown in the figure). Thereafter, the processing performed by the acceleration suppression operation condition determination unit 34 proceeds to step S114.
In step S114, the acceleration suppression operation condition determination unit 34 determines whether or not the condition that the depression amount (operation amount) of the accelerator pedal 32 is equal to or larger than a preset threshold accelerator operation amount (in the drawing). The “accelerator pedal operation determination process” shown in FIG. Here, the process of step S114 is performed based on the depression amount of the accelerator pedal 32 acquired in step S112.

In the present embodiment, as an example, a case will be described in which the threshold accelerator operation amount is set to an operation amount corresponding to 3% of the opening of the accelerator pedal 32. Further, the threshold accelerator operation amount is not limited to an operation amount corresponding to 3% of the opening degree of the accelerator pedal 32. For example, the threshold accelerator operation amount is changed according to the specifications of the vehicle V such as the braking performance of the vehicle V. May be.
When it is determined in step S114 that the condition that the depression amount (operation amount) of the accelerator pedal 32 is equal to or greater than the threshold accelerator operation amount is satisfied (“Yes” shown in the drawing), the acceleration suppression operation condition determination unit 34 The processing to be performed proceeds to step S116.

On the other hand, if it is determined in step S114 that the condition that the depression amount (operation amount) of the accelerator pedal 32 is equal to or greater than the threshold accelerator operation amount is not satisfied ("No" in the drawing), the acceleration suppression operation condition determination unit The processing performed by 34 proceeds to step S120.
In step S116, the acceleration suppression operation condition determination unit 34 acquires information for determining whether or not the vehicle V enters the parking area ("parking area entry determination information acquisition process" shown in the figure). Do. Thereafter, the processing performed by the acceleration suppression operation condition determination unit 34 proceeds to step S118. In the present embodiment, as an example, whether or not the vehicle V enters the parking area based on the steering angle of the steering wheel 28, the angle formed between the vehicle V and the parking area, and the distance between the vehicle V and the parking area. The case of determining will be described.

Here, a specific example of the process performed in step S116 will be described.
In step S116, the acceleration suppression operation condition determination unit 34 refers to the steering angle signal received from the steering angle calculation unit 10C, and acquires the rotation angle (steering angle) of the steering wheel 28. In addition to this, based on the bird's-eye view image around the vehicle V included in the bird's-eye view image signal received from the surrounding environment recognition information calculation unit 10A, the angle α formed by the vehicle V and the parking area L0, and the vehicle V and the parking area L0. To obtain the distance D.
Here, as shown in FIG. 8, for example, the angle α is an absolute value of an intersection angle between the straight line X and the parking area line L1 and the parking area L0 side.

The straight line X is a straight line in the front-rear direction of the vehicle V that passes through the center of the vehicle V (a straight line extending in the traveling direction). It is a parking area line of the parking area L0 part which becomes parallel or substantially parallel to. Further, the parking area L0 side line is a line on the parking area L0 side that is an extension of L1.
The distance D is, for example, the distance between the center point PF of the front end face of the vehicle V and the center point PP of the entrance L2 of the parking area L0 as shown in FIG. However, the distance D is a negative value after the front end surface of the vehicle V passes through the entrance L2 of the parking area L0. The distance D may be set to zero after the front end surface of the vehicle V passes through the entrance L2 of the parking area L0.

Here, the position on the vehicle V side for specifying the distance D is not limited to the center point PF, and may be, for example, a position set in advance in the vehicle V and a position set in advance in the entrance L2. In this case, the distance D is a distance between a position set in advance in the vehicle V and a position set in advance in the entrance L2.
As described above, in step S116, as information for determining whether or not the vehicle V enters the parking region L0, the steering angle, the angle α between the vehicle V and the parking region L0, the vehicle V and the parking region L0 Get the distance D.

In step S118, the acceleration suppression operation condition determination unit 34 determines whether or not the vehicle V enters the parking area based on the information acquired in step S116 ("parking area entry determination process" shown in the figure). )I do.
If it is determined in step S118 that the vehicle V does not enter the parking area ("No" shown in the figure), the processing performed by the acceleration suppression operation condition determination unit 34 proceeds to step S120.
On the other hand, if it is determined in step S118 that the vehicle V enters the parking area (“Yes” shown in the figure), the processing performed by the acceleration suppression operation condition determination unit 34 proceeds to step S122.

Here, a specific example of the process performed in step S118 will be described.
In step S118, for example, when all of the following three conditions (A1 to A3) are satisfied, it is determined that the vehicle V enters the parking area.
Condition A1. The elapsed time after the steering angle detected in step S116 is equal to or greater than a preset steering angle value (eg, 45 [deg]) is within a preset setup time (eg, 20 [sec]). is there.
Condition A2. The angle α between the vehicle V and the parking area L0 is equal to or less than a preset angle (for example, 40 [deg]).
Condition A3. The distance D between the vehicle V and the parking area L0 is equal to or less than a preset set distance (for example, 3 [m]).

In addition, as a process which judges whether the vehicle V approachs a parking area, you may use the process performed when the parking area approach reliability setting part 38 sets a parking area approach reliability.
Further, the process used to determine whether or not the vehicle V enters the parking area is not limited to the process using a plurality of conditions as described above, but one or more conditions among the above-described three conditions. You may use the process judged by. Moreover, you may use the process which judges whether the vehicle V approachs into a parking area | region using the vehicle speed of the vehicle V. FIG.
In step S120, the acceleration suppression operation condition determination unit 34 generates an acceleration suppression operation condition determination result signal as an information signal including a determination result that the acceleration suppression control operation condition is not satisfied ("Acceleration suppression operation condition shown in the figure"). Not established ”). Thereafter, the processing performed by the acceleration suppression operation condition determination unit 34 proceeds to step S124.

In step S122, the acceleration suppression operation condition determination unit 34 generates an acceleration suppression operation condition determination result signal as an information signal including a determination result that the acceleration suppression control operation condition is satisfied ("acceleration suppression operation condition shown in the figure"). Established)). Thereafter, the processing performed by the acceleration suppression operation condition determination unit 34 proceeds to step S124.
In step S124, the acceleration suppression operation condition determination unit 34 outputs the acceleration suppression operation condition determination result signal generated in step S120 or step S122 to the acceleration suppression command value calculation unit 10J ("Acceleration suppression operation shown in the figure"). “Condition judgment result output”). Thereafter, a series of processing ends (END).

Process Performed by Parking Area Confidence Setting Unit 36 The process in which the parking area certainty setting unit 36 sets the parking area certainty will be described using FIGS. 9 to 19 with reference to FIGS. 1 to 8.
As shown in FIG. 9, when the parking area certainty setting unit 36 starts the process (START), first, the process proceeds to step S200.
In step S200, the parking area certainty level setting unit 36 performs a process of setting the parking area certainty level to the lowest value (level 0) ("set to level 0" shown in the drawing). Then, the process which the parking area reliability setting part 36 performs transfers to step S202.

In step S202, the parking area certainty setting unit 36 acquires a bird's-eye view around the vehicle V included in the bird's-eye image signal received from the surrounding environment recognition information calculation unit 10A ("Ambient image acquisition shown in the figure"). Process "). Then, the process which the parking area reliability setting part 36 performs transfers to step S204.
In step S204, the parking area certainty setting unit 36 determines whether or not the determination element is suitable for the conditions as the parking area, based on the parking area determination element included in the overhead image acquired in step S202. Processing (“parking area conformity condition determination processing” shown in the figure) is performed. Thereafter, the process performed by the parking area certainty setting unit 36 proceeds to step S206.

Here, a specific example of the process performed in step S204 will be described with reference to FIG.
As shown in FIG. 10, when the parking area certainty level setting unit 36 starts processing (START), first, the process proceeds to step S2000.
In step S2000, the parking area certainty setting unit 36 determines whether or not the pair conforms to the conditions of the end portions constituting the parking area based on the end candidate pairs included in the overhead view image (FIG. The “edge determination process” shown in FIG.

Hereinafter, a specific example of the edge determination process will be described with reference to FIGS. 11 to 14.
First, the edge candidate extraction unit 36a extracts edge candidates located on the road surface from the overhead view image in the vehicle traveling direction.
As shown in FIG. 11A, when detecting the end portions Pm and Pn, the acquired overhead image is scanned in the horizontal direction. When scanning an image, for example, a monochrome image obtained by binarizing a captured image is used. Since the end portion of the parking area is displayed in white or the like that is sufficiently brighter than the road surface, the brightness is higher than that of the road surface. For this reason, as shown in FIG. 11B, a positive edge in which the luminance is rapidly increased is detected at the boundary portion where the road surface changes to the end portion. On the other hand, at the boundary portion where the edge portion changes to the road surface, a negative edge where the luminance decreases rapidly is detected. In FIG. 11B, the plus edge is indicated by a sign “E +” and the minus edge is indicated by a sign “E−”.

FIG. 11B is a graph showing the luminance change of the pixels in the image when scanning from left to right is performed. Note that (1) to (4) in FIG. 11 (b) and (1) to (4) in FIG. 11 (a) correspond to each other.
Here, as shown to Fig.5 (a)-(q), a parking area generally partitions a road surface area | region with a linear parking area line. However, as shown in FIGS. 6A to 6B, for example, there is a parking area in which a rectangular road surface area is partitioned by only four end portions (patterns having a predetermined shape) arranged at four corners. The example of Fig.11 (a) has shown the parking area | region comprised from four L-shaped edge parts shown in Fig.6 (a).
As a parking area composed of four end portions, there is, for example, a parking region composed of four U-shaped end portions shown in FIG.

In this embodiment, in order to detect the parking area illustrated in FIGS. 6A to 6B, the edge candidate extraction unit 36a first detects an image element estimated as an edge candidate from the overhead image. To do. Then, the edge candidate extraction unit 36a determines whether or not the detected image element meets the condition as the edge candidate.
In the process of recognizing the image element estimated as the edge candidate, the plus edge (E +) and minus edge (E +) and minus edge (E−) are detected from the state in which neither the plus edge (E +) nor the minus edge (E−) is detected in the scanning direction. A pair of adjacent edges are detected in the order of E-). Such an edge is continuously detected in a direction perpendicular to the scanning direction at the proximity position. In addition, as shown in the end portions Pm and Pn in FIG. 11A, the end portions constituting the parking area of only the end portions (in the following description, may be described as “single end portions”), An edge is not detected at a position that is shorter than the end of the parking area line. Therefore, in addition to continuous edge detection, it is determined that there is an image element that is estimated to be a single end by detecting a state without an edge at a relatively short position. Note that the number of times of continuous detection is set in advance from the actual length of the single end, the scanning interval in the orthogonal direction, and the like.

For example, as shown in (1) in FIG. 11B, the edges as shown in (2) to (3) in FIG. 11B are continuously set a predetermined number of times from the state in which no edge is detected. To detect. At this time, as shown in (4) of FIG. 11B, when an edge is not detected halfway, an image area in which the edge is continuously detected is designated as an edge candidate (single edge). Detected as an image element estimated to be a part candidate).
In the present embodiment, for example, even when most of the parking area line constituting the parking area disappears due to aging deterioration or the like and only the end portion remains, this image area is changed to the end portion. It detects as an image element estimated as a candidate.

Next, when the state of the detected image element satisfies all of the following two conditions (B1 to B2), for example, the edge candidate extraction unit 36a extracts the image element as an edge candidate. .
Condition B1. The width of the image element is equal to or larger than a preset setting width (for example, a width corresponding to 10 [cm] on the road surface).
Condition B2. The length of the image element is not less than a preset set length (for example, a length corresponding to 30 [cm] on the road surface).
In the present embodiment, image elements that satisfy all of the conditions B1 and B2 are extracted as edge candidates.
In the case of a U-shaped end, two pairs of edges are detected at short intervals (intervals less than a preset interval threshold) from the middle. Thus, two pairs of edges detected at short intervals are extracted as one end candidate.

Next, the parking area detection unit 36d extracts two edge candidates that are adjacent within a preset interval range as edge candidate pairs from the edge candidates extracted by the edge candidate extraction unit 36a.
In the present embodiment, a pair of end candidates satisfying a preset end pair extraction condition from the extracted end candidates is described as an end candidate pair (in the following description, “end candidate pair”). Extract as).
In the present embodiment, as shown in FIG. 12A, a range of length d1 (for example, 1.8 [m]) to d2 (for example, 3.0 [m]) is set in advance as the interval range. ing.
That is, as shown in FIG. 12A, when the distance ds1 between two edge candidates is within the range of d1 to d2 (d1 <ds1 <d2), the distance between the two edge candidates. Is within the preset interval range.

On the other hand, for example, as shown in FIG. 12B, when the distance ds1 between the two end candidates is longer than d2 (d2 <ds1), the interval between the two end candidates is set in advance. It is determined that it is not within the range (outside the interval range).
Also, for example, as shown in FIG. 12C, when the distance ds1 between the two end candidates is shorter than d2 (ds1 <d1), the interval between the two end candidates is set in advance. It is determined that it is not within the range (outside the interval range).
And the parking area | region detection part 36d extracts the end part candidate group determined that the space | interval of two edge part candidates is within the preset space | interval range among several edge part candidates as an edge part pair. .

Next, a process is performed to determine whether or not the combination of the shapes of the extracted edge candidate pairs is compatible with a predefined shape combination.
Specifically, in the present embodiment, the shape of each end of the end candidate pair is a shape defined in advance as a shape pattern of the end, and is defined in advance in the case of a combination of ends having the same shape. It is determined that the combination of the shapes is suitable. The edge shape is determined by, for example, pattern matching using a template corresponding to each shape pattern.
In addition, it is good also as a structure determined not only as the combination of the edge parts of the same shape but the combination of a different shape if it is a combination of the actual edge part shape.

For example, when the combination of the shapes of the end candidate pairs is an L-shaped end combination as shown in FIG. 13 (a), or a U-shaped end as shown in FIG. 13 (b). In the case of a combination of parts, it is determined that the combination of shapes is predefined.
Further, for example, the combinations of the shapes of the edge candidate pairs are different from each other, such as combinations of L-shaped ends and U-shaped ends, as shown in FIGS. 13C and 13D. In the case of the combination of the end portions of the shape, it is determined that the shape combination does not conform to the predefined shape combination.

Next, a process of determining whether or not the deviation of the orientation of the edge candidate pair is equal to or less than a preset deviation threshold value is performed.
For example, in the case of an L-shaped end candidate pair, there is no deviation in the direction of both in the positional relationship shown in FIG. In the present embodiment, a deviation in the direction of the L-shaped end (angle θd) is detected on the basis of the direction of both shown in FIG. 14A, and this deviation angle θd is set to a preset deviation threshold (for example, 5 °) or less.
For example, as shown in FIGS. 14B and 14C, when one of the end candidate pairs has a deviation, it is determined whether or not the deviation angle θd is equal to or smaller than a deviation threshold. In the example of FIG. 14B, the deviation angle θd is larger than the deviation threshold, and in the example of FIG. 14C, the deviation angle θd is less than the deviation threshold. Therefore, it is determined that the deviation of the direction of the end candidate pair in FIG. 14B is larger than the deviation threshold, and the deviation of the direction of the end candidate pair in FIG. Is done.

Also, for example, in the case of a U-shaped end candidate pair, there is no deviation in the direction of both in the positional relationship shown in FIG. Also in this case, a deviation (angle θd) in the direction of the U-shaped end portion is detected on the basis of both directions shown in FIG. 3 [°]) or less.
For example, as shown in FIGS. 14E and 14F, when there is a deviation in both of the end candidate pairs, it is determined whether or not the deviation angle θd is equal to or smaller than a deviation threshold. In the example of FIG. 14E, the deviation angle θd between both ends is larger than the deviation threshold. Therefore, it is determined that the deviation in the direction of the end candidate pair in FIG. 14E is larger than the deviation threshold. On the other hand, in the example of FIG. 14C, the deviation angle θd at the left end is larger than the deviation threshold, but the deviation angle θd at the right end is equal to or less than the deviation threshold. Also in this case, it is determined that the deviation of the direction of the end candidate pair in FIG. 14F is larger than the deviation threshold. That is, if there is a deviation larger than the deviation threshold in one of the edge candidate pairs, it is determined that the deviation in the direction of the edge candidate pair is larger than the deviation threshold.

Next, processing for setting a preset edge determination flag is performed. The initial value of the edge determination flag is “OFF”.
Specifically, when the combination of the shape of the edge candidate pair is defined in advance and the deviation of the direction of the edge candidate pair is equal to or less than the deviation threshold, the condition of the edge constituting the parking area is satisfied. The end determination flag is set to ON. That is, it is assumed that a parking area composed of such end candidate pairs has been detected. Thereafter, the process performed by the parking area certainty setting unit 36 proceeds to step S2010.

On the other hand, if any one of the combination of the shape of the edge candidate pair, the interval between the edge candidate pairs, and the deviation of the direction of the edge candidate pair does not conform to the conditions of the edges constituting the parking area, Set the part determination flag to OFF. That is, it is assumed that the end candidate pair does not constitute a parking area. Thereafter, the process performed by the parking area certainty setting unit 36 proceeds to step S2010.
Returning to FIG. 10, in step S2010, in the parking area certainty setting unit 36, based on the pair of parking area line candidates included in the overhead view image, whether or not the pair meets the conditions of the lines constituting the parking area. Is performed ("area line determination process" shown in the figure).
First, the line extraction unit 36b acquires a line marked on the road surface from the overhead image acquired in step S202.

Next, for example, when the acquired line state satisfies all of the following three conditions (C1 to C3), the line is extracted as a parking area line candidate.
Condition C1. If the marked line on the road surface has a broken part, the broken part is a part where the marked line is faint (for example, a part having a lower clarity than the line and a higher clarity than the road surface). ).
Condition C2. The width of the line marked on the road surface is not less than a preset setting width (for example, 10 [cm]).
Condition C3. The length of the line marked on the road surface is greater than or equal to a preset set line length (for example, 2.5 [m]).
Here, the parking area line candidate is a candidate for a line constituting the parking area, and is a line (white line or the like) marked on the road surface.

Hereinafter, a method for extracting a parking area line candidate that is a determination element will be described.
As shown in FIG. 15A, when the parking area line candidates Lm and Ln are detected, scanning in the horizontal direction is performed in the area indicating the captured image. When scanning an image, for example, a monochrome image obtained by binarizing a captured image is used. FIG. 15A shows a captured image. Since the parking area line is displayed in white or the like sufficiently brighter than the road surface, the brightness is higher than that of the road surface. For this reason, as shown in FIG. 15B, a positive edge in which the luminance is rapidly increased is detected at the boundary portion where the road surface changes to the parking area line. FIG. 15B is a graph showing the luminance change of the pixels in the image when scanning from left to right, and FIG. 15C is the same as FIG. 15A. FIG. Further, in FIG. 15B, the plus edge is indicated by a sign “E +”, and in FIG. 15C, the plus edge is indicated by a thick solid line with a sign “E +”. In addition, a negative edge at which the luminance decreases rapidly is detected at the boundary portion where the parking area line changes to the road surface. In FIG. 15B, the minus edge is indicated by a sign “E−”, and in FIG. 15C, the minus edge is indicated by a thick dotted line with a sign “E−”. In the process of recognizing the parking area line candidate, the parking area line candidate is detected by detecting a pair of adjacent edges in the order of plus edge (E +) and minus edge (E−) in the scanning direction. Is determined to exist.

Next, in the parking area detection unit 36d, two lines adjacent to each other in the overhead image are selected from the parking area line candidates extracted from the same overhead image (the overhead image corresponding to the individual image) by the line extraction unit 36b. It is specified as one set (in the following description, it may be described as “pairing”). When three or more lines are extracted from the same bird's-eye view image, two or more sets are specified by two adjacent lines for each of the three or more lines.
Next, in the parking area detection unit 36d, it is determined whether or not the paired two lines, which are a pair of parking area line candidates, meet the conditions of the lines constituting the parking area.

Here, referring to FIG. 16, whether or not a pair of parking area line candidates (may be described as “area line candidate pair” in the following description) meets the conditions of the lines constituting the parking area. A specific example of the process for determining whether will be described. In FIG. 16, a region indicating an image captured by the front camera 14 </ b> F in the overhead view image is denoted by reference numeral “PE”.
For example, when the parking area detection unit 36d satisfies all of the following four conditions (D1 to D4) with respect to the two paired lines that are pairs of parking area line candidates, It is determined that the conditions of the lines constituting the parking area are met.
Condition D1. As shown in FIG. 16A, the width WL between two paired lines (in the figure, indicated by the symbols “La” and “Lb”) is a preset pairing width (for example, 2.5 [m]) or less.

Condition D2. As shown in FIG. 16B, the angle (degree of parallelism) formed by the line La and the line Lb is within a preset angle (for example, 3 [°]).
In FIG. 16B, a reference line (a line extending in the vertical direction of the region PE) is indicated by a dotted line with a reference “CLc”, and a central axis of the line La is indicated by a reference “CLa”. The central axis of the line Lb is indicated by a broken line with the sign “CLb”. Further, the inclination angle of the central axis line CLa with respect to the reference line CLc is indicated by a symbol “θa”, and the inclination angle of the central axis line CLb with respect to the reference line CLc is indicated by a reference symbol “θb”.
Therefore, when the conditional expression of | θa−θb | ≦ 3 [°] is satisfied, the condition C2 is satisfied.

Condition D3. As shown in FIG. 16C, a straight line connecting the end of the line La on the vehicle V side (the end on the lower side in the figure) and the end of the line Lb on the vehicle V side is close to the vehicle V. The angle θ formed with the side line L is equal to or larger than a preset set deviation angle (for example, 45 [°]).
Condition D4. As shown in FIG. 16D, the absolute value (| W0−W1 |) of the difference between the width W0 of the line La and the width W1 of the line Lb is a preset line width (for example, 10 [cm]). )
In the process of determining whether or not the four conditions (D1 to D4) described above are satisfied, when the length of at least one of the lines La and Lb is interrupted at about 2 [m], for example, Further, the processing is continued as a line of about 4 [m] obtained by extending a virtual line of about 2 [m].

When the parking area detection unit 36d determines that the area line candidate pair conforms to the conditions of the lines constituting the parking area, the parking area detection unit 36d performs a process of setting a preset area line determination flag to ON (eg, “1”). Do. That is, it is assumed that a parking area composed of such area line candidate pairs has been detected.
The initial value of the area line determination flag is “OFF” (for example, “0”). Then, the process which the parking area reliability setting part 36 performs transfers to step S2020.

On the other hand, when the parking area detection unit 36d determines that the area line candidate pair does not conform to the conditions of the lines constituting the parking area, the parking area detection unit 36d performs a process of setting a preset area line determination flag to OFF. That is, it is assumed that such a region line candidate pair does not constitute a parking region. Then, the process which the parking area reliability setting part 36 performs transfers to step S2020.
Returning to FIG. 10, in step S2020, in the parking area certainty setting unit 36, the edge candidate or the edge candidate pair extracted in step S2000 and the number area included in the overhead image constitute the parking area and Processing for determining whether or not the condition of the number area is met (“first number determination processing” shown in the figure) is performed.

Hereinafter, a specific example of the first numeral determination process will be described with reference to FIGS. 1 to 16 and FIGS. 17 to 20.
As shown in the flowchart of FIG. 17, when the parking area certainty setting unit 36 starts the first number determination process, first, the process proceeds to step S2100.
In step S2100, the parking area certainty setting unit 36 obtains information on the extraction result of the parking area end extracted in step S2000 (in the following description, it may be simply referred to as “end extraction information”). Then, the process proceeds to step S2110. Here, the edge extraction information includes information on whether or not the edge candidate has been extracted, information on the extracted edge candidate, information on whether or not the parking area edge has been detected, and the detected parking area edge. Part information.

In step S2110, the parking area certainty setting unit 36 performs a process of determining whether or not an edge candidate has been extracted based on the edge extraction information acquired in step S2100.
If it is determined in step S2110 that the edge candidate has been extracted ("Yes" shown in the drawing), the process performed by the parking area certainty setting unit 36 proceeds to step S2120.
On the other hand, if it is determined in step S2110 that no end candidate has been extracted ("No" shown in the figure), the process performed by the parking area certainty setting unit 36 proceeds to step S2195.
In step S2120, the number area extraction unit 36c performs a process of extracting a number area (“number area extraction process” shown in the figure). Thereafter, the process proceeds to step S2130.

Hereinafter, a specific example of the number area extraction process in step S2120 will be described with reference to FIG.
As the positional relationship between the parking area and the identification information, for example, as shown in FIGS. 18C to 18D, the identification information is located inside the two parking area ends, FIG. As shown in g) to (h), the position in the left-right direction is the same as that shown in FIGS. 18 (c) to 18 (d). .

  Based on this, in the present embodiment, as shown in FIG. 19, in the number area extraction unit 36 c, an imaginary line orthogonal to the perspective direction through the near end of the extracted edge candidate in the overhead image PE ( A region Ar1 in which the distance in the depth direction and the front direction is within the distance ds2 is set from the broken line in the drawing. Then, as shown in FIG. 20A, the edge detection luminance threshold value is set to a value lower than the threshold values of the other regions for the pixels in the region Ar1. In addition, as shown in FIG. 19, a region Ar2 between the distances ds2 to ds3 in the depth direction and the near direction is set. Then, as shown in FIG. 20A, a luminance threshold value that increases as the distance increases from ds2 to ds3 is set for the pixels in the region Ar2, and after the distance ds3, the value at ds3 is constant. Set the brightness threshold.

That is, in the present embodiment, an edge is easily detected for an area range (area Ar1) in which a numerical area is highly likely to exist for the extracted edge candidate.
For example, in the example of FIG. 11A, a number area including the number “40” is located inside the two front ends Pmf and Pnf. The numbers marked on the road surface are displayed in white or the like that is sufficiently bright as compared with the road surface in the same manner as the end portions. Therefore, as shown in (3) of FIG. 11B, an edge corresponding to the number shape is detected. The luminance threshold value used for edge detection is set to a value lower than other distance ranges in the area Ar1 shown in FIG. 11A, as shown in FIG. 20A.
When the number area extraction unit 36c detects a set of edges, the number area extraction unit 36c then sets in advance an image area corresponding to the set of edges (in the following description, it may be described as “edge set image”). Perform template matching with the template image. Here, the template image includes an image of a symbol including numbers, characters, etc. constituting identification information of an existing parking area.

In template matching, it is determined whether or not the degree of matching (matching rate) between the edge aggregate image and the template image is equal to or higher than a preset matching threshold. At this time, in this embodiment, as shown in FIG. 20B, the matching threshold value is set to a value lower than the threshold values of the other regions for the region Ar1 shown in FIG. In addition, in the area Ar2, a matching threshold value that increases as the distance becomes longer is set between the distances ds2 and ds3, and a matching threshold value that is constant at a value in ds3 is set after the distance ds3.
That is, for the extracted end candidate, the number area is easily extracted for the area range (area Ar1) where the number area is highly likely to exist.
The number area extraction unit 36c extracts the edge set image determined to be equal to or higher than the matching threshold as a number area.

Returning to FIG. 16, in step S <b> 2130, the parking area detection unit 36 d performs a process of determining whether or not a numeric area has been extracted by the numeric area extraction process in step S <b> 2120.
If it is determined in step S2130 that the numeric area has not been extracted ("No" shown in the figure), the process performed by the parking area certainty setting unit 36 proceeds to step S2190.
On the other hand, if it is determined in step S2130 that the numeric area has been extracted (“Yes” shown in the figure), the process performed by the parking area certainty setting unit 36 proceeds to step S2140.

In step S2140, the parking area detection unit 36d performs a process of determining whether or not the numbers included in the extracted number area satisfy the conditions of the numbers used for the parking area.
If it is determined in step S2140 that the conditions of the numbers used for the parking area are satisfied (“Yes” shown in the figure), the process performed by the parking area certainty setting unit 36 proceeds to step S2150.
On the other hand, if it is determined in step S2140 that the numerical condition used for the parking area is not satisfied ("No" shown in the drawing), the process performed by the parking area certainty setting unit 36 proceeds to step S2195.

Hereinafter, a specific example of the process performed in step S2140 will be described with reference to FIGS.
In public roads, such as the speed limit shown in FIG. 21 (a), the school zone time zone shown in FIG. 21 (b), and the bus time zone shown in FIG. There is a sign.
However, since the numbers marked on public roads are assumed to be viewed by the driver from the traveling vehicle, some numbers are marked vertically (particularly speed limit). Based on this, in the present embodiment, as shown in FIG. 22A, the ratio of the vertical width dv and the horizontal width dh of the number area NA (the aspect ratio (dv: dh)) is a preset aspect ratio. It is determined whether it is within the range.

For example, in the indication of the speed limit, the vertical and horizontal dimensions are defined such that the vertical width is 500 [cm] and the horizontal width (two digits) is 120 [cm]. Therefore, the aspect ratio of the numbers (two digits) used for indicating the speed limit is “25: 6”. Therefore, for example, when a two-digit number is included in the number area, the aspect ratio “dv: dh” of the two-digit number is excluded so that the aspect ratio “dv: dh” is “25: 6” is excluded. Set the range. In addition, the range of the aspect ratio is set so as not to include the aspect ratio of numbers that exist only on public roads.
Therefore, based on the aspect ratio of the numbers used as the identification number of the existing parking area, the range from the lowest value to the maximum value is set as the aspect ratio range.

Not only the range of the aspect ratio, but also set the aspect ratio of the numbers that are displayed only on public roads in advance, and determine whether it matches this aspect ratio or whether it is within the preset error range. When it is determined that the numbers match or fall within an error range, the number included in the number area NA may be determined to be a number marked on a public road.
In the example shown in FIG. 22 (a), the numerical area NA is set to a range slightly wider than the vertical and horizontal widths of the numbers included in the area NA. For example, the range may be set to another range as long as the error in determination is within the allowable range.

When the parking area detection unit 36d determines that the aspect ratio is within the range of the aspect ratio, the parking area detection unit 36d sets a preset aspect ratio flag to ON (eg, “1”). Set to OFF (for example, “0”). Here, the aspect ratio flag is a flag indicating that it is within the range of the aspect ratio when in the ON state and is not within the range of the aspect ratio when in the OFF state.
After setting the aspect ratio flag, the parking area detection unit 36d next determines whether or not the size (for example, area) of the numeric area NA is equal to or larger than a preset size threshold.

  Here, the dimensions of the numbers on the road markings are, for example, 500 [cm] for the vertical width and 120 [cm] for the horizontal width (50 [cm] in the case of one digit), as exemplified by the speed limit. The size is larger than the size of the symbol string used as the area identification information (for example, in the case of a two-digit symbol string, the vertical width is 30 [cm] and the horizontal width is 50 [cm]). The same can be said for bus-only times and school zone times. For example, the dimension of a number (one digit) corresponding to a short hand of a clock used for indicating a time dedicated to a bus has a vertical width of 150 [cm], a horizontal width of 70 [cm], and a number corresponding to a long hand of a clock (2 The size of the digit) is specified as 75 [cm] in the vertical width and 85 [cm] in the horizontal width (40 [cm] in the case of one digit).

When the parking area detection unit 36d determines that the size (area) of the numeric area NA is not equal to or larger than the size threshold, the parking area detection unit 36d sets a preset size flag to ON (eg, “1”). On the other hand, if it is determined that the size is greater than or equal to the size threshold, the size flag is set to OFF (eg, “0”).
Here, the size flag indicates that the number included in the numeric area NA is the size used in the parking lot when in the ON state, and the number included in the numeric area NA is used in the parking lot when in the OFF state. It is a flag indicating that there is no size.
When the parking area detection unit 36d determines that both the aspect ratio flag and the size flag are in the ON state, the parking area detection unit 36d determines that the numbers included in the number area NA satisfy the conditions of the numbers used in the parking area.

On the other hand, when the parking area detection unit 36d determines that either the aspect ratio flag or the size flag is in the OFF state, the parking area detection unit 36d determines that the number included in the number area NA does not satisfy the number condition used for the parking area. To do.
In step S2150, the parking area detection unit 36d performs a process of determining whether or not the distance between the extracted end candidate and the extracted number area NA is within a preset first distance range.
If it is determined in step S2150 that the distance between the edge candidate and the numeric area NA is within the first distance range (“Yes” shown in the drawing), the process performed by the parking area certainty setting unit 36 is step S2160. Migrate to
On the other hand, if it is determined in step S2150 that the distance between the end candidate and the number area is not within the first distance range ("No" shown in the drawing), the process performed by the parking area certainty setting unit 36 is performed in step S2150. The process proceeds to S2195.

Hereinafter, a specific example of the process performed in step S2150 will be described with reference to FIGS.
As shown in FIG. 22B, when only one edge candidate is extracted, a vertical line (dotted line in the figure) passing through the middle point in the left-right direction of the one edge candidate Pnf and a numeric area It is determined whether or not the distance ds4 between the vertical line passing through the midpoint in the left-right direction of NA (the one-dot chain line in the drawing) is within a preset first distance range. Note that FIG. 22B illustrates the case where the right-side front end candidate Pnf is extracted, but, of course, only the left-side front end candidate Pmf may be extracted. Also in this case, the distance is determined in the same manner as the edge candidate Pnf.

When the distance ds4 is within the first distance range (for example, a range of 0.8 to 1.2 [m]), the parking area detection unit 36d determines that the distance between the end candidate Pnf and the number area NA is the first. It is determined that the distance is within the range. On the other hand, when the distance ds4 is not within the first distance range, it is determined that the distance between the end candidate Pnf and the numeric area NA is not within the first distance range.
On the other hand, when the edge candidate pair has been extracted, the parking area detection unit 36d, as shown in FIG. 22 (c), sets the distance between the edge candidate pairs Pmf and Pnf and the vicinity thereof. If there is a numerical area NA, the distance is determined as follows. That is, between the vertical line (broken line in the figure) passing through the middle point in the horizontal direction of each of the edge candidate pairs Pmf and Pnf and the vertical line (dotted line in the figure) passing through the middle point in the horizontal direction of the number area NA. It is determined whether the distances ds4 and ds5 are within the first distance range.

The parking area detection unit 36d determines that the distance between the end candidate pairs Pmf and Pnf and the numeric area NA is within the first distance range when both the distances ds4 and ds5 are within the first distance range. On the other hand, when at least one of the distances ds4 and ds5 is not within the first distance range, it is determined that the distance between the end candidate pairs Pmf and Pnf and the numeric area NA is not within the first distance range.
In step S2160, in the parking area detection unit 36d, the deviation of the extracted number area NA from the preset reference direction and the deviation of the extracted end candidate from the preset reference direction are set as the preset deviation. Processing for determining whether or not the threshold value (angle threshold value) or less is performed.

In step S2160, the deviation of the extracted number area NA from the reference direction and the deviation of the extracted edge candidate from the reference direction are both equal to or less than the deviation threshold (“Yes” shown in the drawing). When it determines, the process which the parking area reliability setting part 36 performs transfers to step S2170.
On the other hand, in step S2160, any one of the deviation from the reference direction of the extracted number area NA and the deviation from the reference direction of the extracted edge candidate is not less than the deviation threshold (shown in the figure). When it determines with "No"), the process which the parking area reliability setting part 36 performs transfers to step S2195.

Hereinafter, a specific example of the process performed in step S2160 will be described with reference to FIG.
In this embodiment, as shown in FIG. 22 (d), the deviation from the reference direction of the numerical area NA is defined as the angle θn1 with the reference direction as the direction of the numerical area NA in FIGS. 22 (b) and 22 (c). And set. In addition, as shown in FIG. 22 (d), the deviation from the reference direction of the end portions Pmf and Pnf is determined by using the angle θn2 and the reference direction as the direction of the end candidate candidates Pmf and Pnf in FIG. Set to θn3.
When the angle θn1 and at least one of θn2 to θn3 are equal to or smaller than the deviation threshold, the parking area detection unit 36d detects the deviation from the reference direction of the extracted numerical area NA and the extracted end candidate reference. It is determined that any deviation from the direction is less than the deviation threshold.

On the other hand, the parking area detection unit 36d detects the deviation from the reference direction of the extracted number area NA and the reference direction of the extracted end candidate when at least θn1 of the angles θn1 to θn3 is not less than the deviation threshold. It is determined that all the deviations from are not less than the deviation threshold.
In step S2170, when an end candidate pair is extracted in the parking area detection unit 36d, the ratio between the width between the end candidate pairs Pmf and Pnf and the horizontal width of the number area NA is set in advance. A process for determining whether or not the ratio is equal to or greater than the ratio threshold is performed.
If it is determined in step S2170 that the width between the edge candidates and the horizontal width of the number area NA are equal to or greater than the ratio threshold value, or only one edge candidate has been extracted (“Yes” shown in the drawing). In this case, the process performed by the parking area certainty factor setting unit 36 proceeds to step S2180.
On the other hand, if it is determined in step S2170 that the width between the edge candidates and the horizontal width of the number area NA are not equal to or greater than the ratio threshold value ("No" in the figure), the process performed by the parking area certainty setting unit 36 is as follows. Then, the process proceeds to step S2195.

Hereinafter, a specific example of the process performed in step S2170 will be described with reference to FIG.
In the present embodiment, as shown in FIG. 22E, the horizontal width of the number area NA is ds6, and the width between the end candidate pairs Pmf and Pnf is ds7. Then, it is determined whether or not the ratio “ds6 / ds7” is equal to or greater than a preset ratio threshold.
For example, the maximum value of the horizontal width (2 digits) of the numeric area NA is set to 60 [cm], the maximum value of the horizontal width range of the parking area is set to 3 [m], and these ratios “60/300 = 1/5 = 0. .2 ”is set as the ratio threshold.
In this case, the parking area detection unit 36d determines whether “ds6 / ds7” is equal to or greater than “0.2”. If it is determined that the width is greater than or equal to “0.2”, it is determined that the width between the edge candidates and the horizontal width of the number area NA are equal to or greater than the ratio threshold value. It is determined that the width between candidates and the horizontal width of the number area NA are not greater than or equal to the ratio threshold.

In step S2180, the parking area detection unit 36d includes an image area including a symbol other than a numeral within a second distance range set in advance from the numeral area NA (in the following description, it may be described as “non-numeric area”). The process of determining whether or not exists.
If it is determined in step S2180 that there is a non-numeric area within the second distance range from the numeric area NA (“Yes” shown in the figure), the process performed by the parking area certainty setting unit 36 proceeds to step S2195. .
On the other hand, if it is determined in step S2180 that there is no non-numeric area within the second distance range from the numeric area NA ("No" shown in the figure), the process performed by the parking area certainty setting unit 36 is step S2190. Migrate to

Hereinafter, a specific example of the process performed in step S2180 will be described with reference to FIG.
In the present embodiment, as the area within the second distance range from the numerical area NA, as shown in FIG. 22 (f), an area surrounding the left and right ends and the upper end of the numerical area NA is preset as a peripheral area Ar3. ing. In the peripheral area Ar3, the distance ds9 in the left-right direction from the left and right ends of the number area NA is set to 1 [m], for example, and the distance ds8 in the depth direction from the upper end of the number area NA is set to 2 [m], for example. Is set.
Here, when the number included in the number area NA is a number (identification number) as identification information of the parking area, there are many cases where there are no symbols such as characters or character strings other than numbers around the number. . Therefore, in this embodiment, when a symbol other than a number (for example, a character string of a school zone) exists around the extracted number area, the number area is excluded from candidates for the number area constituting the parking area. .

However, there are exceptions such as parking areas where characters are present around the identification number, such as “light” characters in parking areas for light vehicles. For this reason, the numerical areas constituting such a parking area are set in advance so as not to be excluded from the candidates for the numerical areas constituting the parking area.
The parking area detection unit 36d determines whether or not there is an image area including symbols other than numerals such as letters of “school zone” in the surrounding area Ar3. Is determined to have a non-numeric area. On the other hand, if the parking area detection unit 36d determines that there is no image area including symbols other than numerals such as characters of “school zone” in the surrounding area Ar3, a non-number area exists around the number area NA. Judge that not.

In step S2190, the parking area detection unit 36d performs processing for setting a preset first number determination flag to ON. That is, it is assumed that a parking area composed of the end candidate and the numeric area NA is detected. Thereafter, the series of processes is terminated and the process returns to the original process.
In step S2195, the parking area detection unit 36d performs a process of setting a preset first number determination flag to OFF. Thereafter, the series of processes is terminated and the process returns to the original process.
Returning to FIG. 10, in step S2030, based on the parking area line candidate or area line candidate pair detected in step S2010, the parking area line candidate and the number area included in the overhead image constitute a parking area line. And a process for determining whether or not the condition of the number area is met ("second number determination process" shown in the figure). Thereafter, the series of processes is terminated and the process returns to the original process.

Hereinafter, a specific example of the second numeral determination process will be described with reference to FIGS. 1 to 22 and FIGS. 23 to 26.
As shown in FIG. 23, when the parking area certainty setting unit 36 starts the second number determination process, first, the process proceeds to step S2200.
In step S2200, the parking area certainty setting unit 36 acquires information on the extraction result of the parking area line extracted in step S2010 (in the following description, it may be simply referred to as “area line extraction information”). The process proceeds to step S2210. Here, the area line extraction information includes information on whether or not a parking area line candidate has been extracted, information on the extracted parking area line candidate, information on whether or not a parking area line has been detected, and a detected parking area. Line information.

In step S2210, the parking area certainty setting unit 36 performs a process of determining whether or not a parking area line candidate has been extracted based on the information acquired in step S2200.
If it is determined in step S2210 that a parking area line candidate has been extracted (“Yes” shown in the drawing), the process performed by the parking area certainty setting unit 36 proceeds to step S2220.
On the other hand, if it is determined in step S2210 that no parking area line candidate has been extracted ("No" shown in the figure), the process performed by the parking area certainty setting unit 36 proceeds to step S2295.
In step S2220, the number area extraction unit 36c performs a process of extracting a number area (“number area extraction process” shown in the figure). Thereafter, the process proceeds to step S2230.

Hereinafter, a specific example of the process performed in step S2220 will be described with reference to FIGS. 11 and 18 to 20 and FIG.
As the positional relationship between the parking area and the identification information, for example, as shown in FIGS. 18A to 18B, the identification information is positioned inside two straight lines (vertical lines), right and left The position in the direction is the same as in FIGS. 18A to 18B, but there is a case where the identification information is located on the near side (lower side in plan view) with respect to the two vertical lines.
Based on this, in this embodiment, as shown in FIG. 24, in the number area extraction unit 36c, the virtual line orthogonal to the perspective direction passes through the near end of the extracted parking area line candidate in the overhead image PE. From the (broken line in the figure), a region Ar1 in which the distance in the depth direction and the near direction is within the distance ds2 is set. Then, as shown in FIG. 20A, the brightness threshold for edge detection is set to a value lower than the thresholds of the other areas in the area Ar1. In addition, as shown in FIG. 24, an area Ar2 is set between the distance ds2 in the depth direction and the near direction to ds3. Then, as shown in FIG. 20A, a luminance threshold value that increases as the distance becomes longer is set between ds2 and ds3, and a luminance threshold value that is constant at the value of ds3 is set after the distance ds3.

That is, with respect to the extracted parking area line candidate, an edge is easily detected in an area range (area Ar1) where there is a high possibility that a numeric area exists.
When detecting the edge set image, the number area extraction unit 36c performs template matching with a preset template image on the edge set image.
Then, for the area Ar1 shown in FIG. 24, as shown in FIG. 20B, the matching threshold value is set to a value lower than the threshold values of the other areas. In addition, in the area Ar2, a matching threshold value that increases as the distance becomes longer is set between the distances ds2 and ds3, and a matching threshold value that is constant at a value in ds3 is set after the distance ds3.

That is, for the extracted parking area line candidate, the number area is likely to be extracted for the area range (area Ar1) where there is a high possibility that the number area exists.
The number area extraction unit 36c extracts the edge set image determined to be equal to or higher than the matching threshold as a number area.
Here, the processing in steps S2230 to S2240 is the same as the processing in steps S2130 to S2140, and a description thereof will be omitted.
In step S2250, the parking area detection unit 36d performs a process of determining whether or not the distance between the extracted parking area line candidate and the extracted number area NA is within a preset first distance range.

In step S2250, when it is determined that the distance between the extracted parking area line candidate and the extracted number area NA is within the first distance range ("Yes" shown in the drawing), the parking area certainty setting unit 36 performs it. The processing moves to step S2260.
On the other hand, if it is determined in step S2250 that the distance between the extracted parking area line candidate and the extracted number area is not within the first distance range ("No" shown in the figure), the parking area certainty setting unit 36 The processing to be performed moves to step S2295.

Hereinafter, a specific example of the process performed in step S2150 will be described with reference to FIGS.
As shown in FIG. 25 (a), when only one parking area line candidate is extracted, the perspective direction passes through the middle point in the left-right direction at the front end of the one parking area line candidate Ln. A distance ds10 between a vertical line (broken line in the figure) extending in the figure and a vertical line (dashed line in the figure) extending in the perspective direction through the center in the left-right direction of the number area NA is a preset first distance range ( For example, it is determined whether it is within a range of 0.8 to 1.2 [m]. Note that FIG. 25A illustrates the case where the right parking area line Ln is found, but of course, only the left parking area line Lm may be extracted. In this case as well, the distance is determined in the same manner as the parking area line Ln.

When the distance ds10 is within the first distance range, the parking area detection unit 36d determines that the distance between the parking area line candidate and the numeric area NA is within the first distance range. On the other hand, when the distance ds10 is not within the first distance range, it is determined that the distance between the parking area line candidate and the numeric area NA is not within the first distance range.
On the other hand, as shown in FIG. 25 (b), when the area line candidate pair is extracted and there is a numerical area NA between the two parking area line candidates, each parking area line candidate The distances ds10 and ds11 between the vertical line extending in the perspective direction through the middle point in the left-right direction at the front end portion of the image and the vertical line extending in the perspective direction through the center in the left-right direction of the number area NA are respectively first. It is determined whether or not the distance is within the range.

The parking area detection unit 36d determines that the distance between the parking area line candidate and the numeric area NA is within the first distance range when both the distances ds10 and ds11 are within the first distance range. On the other hand, when at least one of the distances ds10 and ds11 is not within the first distance range, it is determined that the distance between the parking area line candidate and the numeric area NA is not within the first distance range.
In step S2260, in the parking area detection unit 36d, the deviation of the extracted number area NA from the preset reference direction and the deviation of the extracted parking area line candidate from the preset reference direction are set in advance. A process of determining whether or not the deviation threshold value (angle threshold value) or less is performed.

In step S2260, the deviation of the extracted number area NA from the reference direction and the deviation of the extracted parking area line candidate from the reference direction are both equal to or less than the deviation threshold ("Yes" shown in the figure). When it determines with, the process which the parking area reliability setting part 36 performs transfers to step S2270.
On the other hand, in step S2260, any one of the deviation from the reference direction of the extracted numeric area NA and the deviation from the reference direction of the extracted parking area line candidate is not less than the deviation threshold (in the figure). If the determination is “No”, the process performed by the parking area certainty setting unit 36 proceeds to step S2295.

Hereinafter, a specific example of the process performed in step S2260 will be described with reference to FIG.
In this embodiment, as shown in FIG. 25 (c), the deviation from the reference direction of the numerical area NA is defined as the angle θn1 with the reference direction as the direction of the numerical area NA in FIGS. 25 (a) and 25 (b). And set. In addition, as shown in FIG. 25 (c), the deviation from the reference direction of the parking area line candidates Lm and Ln is defined as the direction of the parking area line candidates Lm and Ln in FIG. The angles θn4 and θn5 are set.
When the angle θn1 and at least one of θn4 and θn5 are equal to or less than the deviation threshold, the parking area detection unit 36d detects the deviation of the extracted numerical area NA from the reference direction and the extracted parking area line candidate. It is determined that any deviation from the reference direction is equal to or less than the deviation threshold.

On the other hand, when at least θn1 of the angles θn1, θn4, and θn5 is not less than or equal to the deviation threshold, the parking area detection unit 36d detects the deviation of the extracted numeric area NA from the reference direction and the extracted parking area line candidates. It is determined that any deviation from the reference direction is not less than the deviation threshold.
In step S2010, when region line candidate pairs that satisfy all of the above conditions D1 to D4 are extracted, the determination processing of the angles θn4 to θn5 may be omitted for the region line candidate pairs.

In step S2270, when the area line candidate pair is extracted in the parking area detection unit 36d, the ratio between the width between the area line candidate pairs Lm and Ln and the horizontal width of the number area NA is set in advance. A process of determining whether or not the ratio is equal to or greater than the ratio threshold is performed.
In step S2270, when it is determined that the width between the area line candidate pairs and the horizontal width of the number area NA are equal to or larger than the ratio threshold value, or only one parking area line candidate is extracted (“Yes” shown in the figure). "), The process performed by the parking area certainty setting unit 36 proceeds to step S2280.
On the other hand, if it is determined in step S2270 that the width between the area line candidate pairs and the horizontal width of the numeric area NA are not equal to or greater than the ratio threshold ("No" shown in the drawing), the process performed by the parking area certainty setting unit 36 Shifts to step S2295.

Hereinafter, a specific example of the process performed in step S2270 will be described with reference to FIG.
In the present embodiment, as shown in FIG. 26A, the horizontal width of the numeric area NA is set to ds12, and the width between the area line candidate pairs Lm and Ln is set to ds13. Then, it is determined whether or not this ratio “ds12 / ds13” is equal to or greater than a preset ratio threshold.
Here, the ratio threshold value is set to, for example, “0.2” in the same manner as illustrated in step S2170.
Then, when the ratio “ds12 / ds13” is equal to or greater than “0.2”, the parking area detection unit 36d determines that the width between the area line candidate pairs and the horizontal width of the numeric area NA are equal to or greater than the ratio threshold. To do. On the other hand, when the ratio “ds12 / ds13” is not equal to or greater than “0.2”, it is determined that the width between the area line candidate pairs and the horizontal width of the numeric area NA are not equal to or greater than the ratio threshold.

In step S2280, the parking area detection unit 36d performs a process of determining whether or not a non-numeric area exists within the second distance range from the numeric area NA.
If it is determined in step S2280 that a non-numeric area exists within the second distance range from the numeric area NA (“Yes” shown in the figure), the process performed by the parking area certainty setting unit 36 proceeds to step S2295. .
On the other hand, if it is determined in step S2280 that there is no non-numeric area within the second distance range from the numeric area NA ("No" shown in the figure), the process performed by the parking area certainty setting unit 36 is step S2290. Migrate to

Hereinafter, a specific example of the process performed in step S2280 will be described with reference to FIG.
In the present embodiment, as the area within the second distance range from the numerical area NA, as shown in FIG. 26B, an area surrounding the left and right ends and the upper end of the numerical area NA is preset as a peripheral area Ar3. Yes.
In the peripheral area Ar3, the distance ds9 in the left-right direction from the left and right ends of the number area NA is set to 1 [m], for example, in the depth direction from the upper end of the number area NA, as illustrated in step S2180. The distance ds8 is set to 2 [m], for example.

The parking area detection unit 36d determines whether or not there is an image area including a symbol that does not include a number such as “school zone” in the peripheral area Ar3. It is determined that a non-numeric area exists around. On the other hand, when the parking area detection unit 36d determines that there is no image area that includes a symbol that does not include a number such as the letters “school zone” in the surrounding area Ar3, there is a non-numeric area around the number area NA. Judge that it does not exist.
In step S2290, the parking area detection unit 36d performs a process for setting a preset second number determination flag to ON. That is, it is assumed that a parking area composed of the parking area line candidates and the numeric area NA is detected. Thereafter, the series of processes is terminated and the process returns to the original process.
In step S2295, the parking area detection unit 36d performs a process of setting a preset second number determination flag to OFF. Thereafter, the series of processes is terminated and the process returns to the original process.

Returning to FIG. 9, in step S <b> 206, in the parking area certainty degree calculation unit 36 e, the parking area determination elements included in the overhead image acquired in step S <b> 202 based on the various flags set in step S <b> 202 constitute the parking area. To determine whether or not the condition is met. This processing corresponds to “parking area condition conformance?” Shown in the figure.
In step S206, when it is determined that the extracted determination element of the parking area does not conform to the conditions constituting the parking area ("No" shown in the drawing), the process performed by the parking area certainty setting unit 36 is as follows. The process proceeds to step S200.

Specifically, the parking area certainty calculation unit 36e determines that the extracted parking area when the end determination flag, the area line determination flag, the first number determination flag, and the second number determination flag are all set to OFF. It is determined that the determination element does not conform to the conditions constituting the parking area.
On the other hand, if it is determined in step S206 that the extracted determination element of the parking area is suitable for the conditions constituting the parking area (“Yes” shown in the drawing), the parking area certainty setting unit 36 performs processing. Proceeds to step S208.
Specifically, the parking area certainty degree calculation unit 36e has one of the edge determination flag, the area line determination flag, the first number determination flag, and the second number determination flag set to ON. In addition, it is determined that the extracted determination element of the parking area is suitable for the conditions constituting the parking area.

In step S208, the parking area certainty calculation unit 36e sets the level of the parking area certainty to a level (level 1) that is one level higher than the lowest value (level 0) ("level 1" shown in the figure). Settings ”). Then, the process which the parking area reliability setting part 36 performs transfers to step S210.
In step S210, in the parking area certainty calculation unit 36e, the process of step S206 is continuously adapted from the start of the process of step S206 until the moving distance of the vehicle V reaches the preset set moving distance. The process ("Continuous verification matching?" Shown in the figure) for determining whether or not it is determined is performed. The set movement distance is set in the range of 1.0 to 2.5 [m], for example, according to the specifications of the vehicle V and the forward or backward state. Moreover, the process performed by step S210 is performed with reference to the overhead image signal received from 10 A of surrounding environment recognition information calculating parts, and the vehicle speed calculated value signal received from the own vehicle vehicle speed calculating part 10B, for example.

In step S210, if it is determined that the process of step S206 is not continuously adapted ("No" shown in the figure), the process performed by the parking area certainty setting unit 36 is performed in step S200. Transition.
On the other hand, if it is determined in step S210 that the process of step S206 is continuously adapted (“Yes” shown in the figure), the process performed by the parking area certainty setting unit 36 proceeds to step S212.
Here, in the process performed in step S210, for example, as shown in FIG. 27, the state in which the process in step S206 is determined to be suitable and the state in which the process in step S206 is determined to be incompatible. In accordance with the above, the moving distance of the vehicle V is virtually calculated. In addition, FIG. 27 is a figure which shows the content of the process which the parking area reliability calculation part 36e performs. Further, in FIG. 27, in the area described as “collation state”, the state in which the process in step S206 is determined to be suitable is indicated as “ON”, and the process in step S206 is determined to be incompatible. This state is indicated as “OFF”. Further, in FIG. 27, the travel distance of the vehicle V virtually calculated is indicated as “virtual travel distance”.

As shown in FIG. 27, when the collation state is “ON”, the virtual travel distance increases. On the other hand, when the collation state is “OFF”, the virtual travel distance decreases.
In the present embodiment, as an example, a case will be described in which the slope (increase gain) when the virtual travel distance increases is set larger than the slope (decrease gain) when the virtual travel distance decreases. In other words, if the time when the “verification state” is “ON” and the time when the “verification state” is “OFF” are the same time, the virtual travel distance increases.
Then, when the virtual travel distance reaches the set travel distance without returning to the initial value (shown as “0 [m]” in the figure), it is determined that the process of step S206 is continuously adapted. Judge that

In step S212, the parking area certainty calculation unit 36e sets the level of the parking area certainty to a level (level 2) that is two steps higher than the lowest value (level 0) ("level 2" shown in the figure). Settings ”). Thereafter, the process performed by the parking area certainty level setting unit 36 proceeds to step S214.
In step S214, the parking area certainty setting unit 36 first performs a process of determining whether or not the end determination flag is in an ON state. When it is determined that the end determination flag is in the ON state, the process performed by the parking area certainty setting unit 36 proceeds to step S216.

  On the other hand, when it is determined that the edge determination flag is not in the ON state, a process for determining whether or not the area line determination flag is in the ON state is performed. When it is determined that the area line determination flag is in the ON state, the ends positioned on the same side with respect to the vehicle V with respect to the lines La and Lb in which the processing in step S206 is continuously verified, respectively. Part (near end or far end) is detected. Then, a process of determining whether or not the end portions located on the same side face each other along the direction of the width WL (“approaching near and far end portion?” Shown in the drawing) is performed. The process performed in step S214 is performed with reference to, for example, an overhead image signal received from the ambient environment recognition information calculation unit 10A and a vehicle speed calculation value signal received from the host vehicle vehicle speed calculation unit 10B.

In step S214, when it is determined that the ends positioned on the same side do not face each other along the direction of the width WL ("No" shown in the drawing), the process performed by the parking area certainty setting unit 36 is performed. The process proceeds to step S222.
On the other hand, in step S214, when it is determined that the ends located on the same side face each other along the direction of the width WL ("Yes" shown in the drawing), the parking area certainty setting unit 36 performs the process. The process proceeds to step S216.
In step S216, the parking area certainty calculation unit 36e sets the level of the parking area certainty to a level (level 3) that is three levels higher than the lowest value (level 0) ("level 3" shown in the figure). Settings ”). Then, the process which the parking area reliability setting part 36 performs transfers to step S218.

  In step S218, in the parking area certainty setting unit 36, in the processing in step S214, the end pair that is determined to face each other along the direction of the width WL is further positioned on the other side. Is detected. That is, in the process of step S214, when an end on the side closer to the vehicle V (one side) is detected, an end on the side farther from the vehicle V (the other side) is detected in step S218. . Then, a process of determining whether or not the end portions located on the other side face each other along the direction of the width WL (“end-to-end end-to-end matching?” Shown in the figure) is performed. The process performed in step S218 is performed with reference to, for example, an overhead image signal input from the surrounding environment recognition information calculation unit 10A and a vehicle speed calculation value signal input from the host vehicle vehicle speed calculation unit 10B.

When detecting the ends of the lines La and Lb, for example, a straight end such as the end of the line shown in FIG. 5A and the upper end of the line shown in FIG. All the intersections between the U-shaped end portion such as the portion and the double line and the horizontal line shown in FIG. 5 (o) are processed as one straight end portion. Similarly, there is a gap in the end of a double line such as the upper end of the line shown in FIG. 5H and a U-shaped curve such as the upper end of the line shown in FIG. All the formed end portions are processed as one straight end portion.
Further, when detecting the ends of the lines La and Lb, for example, an inclined double line extending in the vertical direction shown in FIG. 5 (n) and a single straight line extending in the horizontal direction Are not processed (recognized) as end portions. This is because, when detecting the edge, the edge is detected by scanning in the horizontal direction in the region indicating the captured image. Further, for example, the area indicated by the white frame in FIG. 5 (p) indicates an object on the road such as a pillar, and therefore the end of this object is not detected.

In step S218, when it is determined that the end portions located on the other side do not face each other along the direction of the width WL ("No" shown in the drawing), the process performed by the parking area certainty setting unit 36 Proceeds to step S222.
On the other hand, in step S218, when it is determined that the ends located on the other side face each other along the direction of the width WL ("Yes" shown in the drawing), the parking area certainty setting unit 36 The process to be performed moves to step S220.
In step S220, the parking area certainty calculation unit 36e sets the level of the parking area certainty to a level (level 4) that is four levels higher than the lowest value (level 0) ("level 4" shown in the figure). Settings ”). If the process which sets parking area reliability to level 4 is performed in step S220, the process which the parking area reliability setting part 36 performs will transfer to step S222.

Therefore, in the process of setting the parking area certainty level to level 3, among the parking areas shown in FIG. It will be set. Further, in the process of setting the parking area certainty to level 4, among the parking areas shown in FIG. Will be set. Moreover, in the process which sets a parking area reliability to level 4, a parking area reliability is set with respect to the pattern of the parking area of (a) and (b) shown in FIG.
In step S222, the parking area certainty level setting unit 36 determines whether or not the preset condition for the process performed by the parking area certainty level setting unit 36 is satisfied (“end condition satisfied?” Shown in the figure). )I do.

Specifically, for example, based on the shift position signal received from the shift position sensor 20, whether or not the shift position is in the parking ("P") shift position, and the termination condition based on the detection of ignition ON → OFF, etc. It is determined whether or not the above is satisfied. In addition, it may be configured to determine that the end condition is satisfied when the vehicle speed of the vehicle V is equal to or higher than the threshold vehicle speed.
If it is determined in step S222 that the end condition is satisfied, the process performed by the parking area certainty level setting unit 36 is ended (END).

On the other hand, when it is determined in step S222 that the end condition is not satisfied, the parking area certainty setting unit 36 repeatedly performs the determination process until the end condition is satisfied.
The series of processes performed by the parking area certainty setting unit 36 is repeatedly performed every time the start condition is satisfied. For example, it is determined that the start condition is satisfied when the ignition is turned from OFF to ON and the vehicle speed of the vehicle V changes from a state below the threshold vehicle speed to a state above the threshold vehicle speed. In addition, for example, the start condition is satisfied when the vehicle speed of the vehicle V changes from a state where the vehicle speed is greater than or equal to the threshold vehicle speed to a state where the vehicle speed is less than the threshold vehicle speed, or when the vehicle V travels beyond a preset distance in a state where the vehicle speed is less than the threshold vehicle speed Is determined. In addition, for example, when the navigation device 26 detects that the vehicle V has entered the parking lot, it is determined that the start condition is satisfied.

Processing performed by the parking area approach certainty factor setting unit 38 With reference to FIGS. 1 to 27, the parking area approach certainty factor setting unit 38 sets the parking area approach certainty factor with reference to FIGS. 28 and 29. explain.
The parking area approach reliability setting unit 38 performs the process described below for each preset sampling time (for example, 10 [msec]).
As shown in FIG. 28, when the parking area approach certainty setting unit 38 starts the process (START), first, in step S300, a process of detecting the amount of deviation between the predicted locus of the vehicle V and the parking area (in the figure). The “deviation amount detection” shown in FIG. If the process which detects the deviation | shift amount of the anticipated locus | trajectory of the vehicle V and a parking area is performed in step S300, the process which the parking area approach reliability setting part 38 performs will transfer to step S302. In the present embodiment, as an example, a case will be described in which the unit of deviation detected in step S300 is [cm]. Moreover, in this embodiment, the case where the width | variety of a parking area is 2.5 [m] is demonstrated as an example.

  Here, in the process performed in step S300, for example, as shown in FIG. 29, an expected rear wheel trajectory TR of the vehicle V is calculated, and an intersection TP between the calculated expected rear wheel trajectory TR and the entrance L2 of the parking area L0. Is calculated. Further, a distance Lfl between the left parking area line L1l of the parking area L0 and the intersection TP and a distance Lfr between the right parking area line L1r of the parking area L0 and the intersection TP are calculated, and the distance Lfl and the distance Lfr are compared. Then, the longer one of the distance Lfl and the distance Lfr is detected as a deviation amount between the expected rear wheel trajectory TR of the vehicle V and the parking area L0. FIG. 29 is a diagram showing the contents of processing for detecting the amount of deviation between the predicted rear wheel trajectory TR of the vehicle V and the parking area L0.

Further, when calculating the predicted rear wheel trajectory TR of the vehicle V, the center point PR in the vehicle width direction of the right rear wheel WRR and the left rear wheel WRL of the vehicle V is set as a reference point of the vehicle V. . Then, the virtual moving path of the center point PR is calculated using the images captured by the front camera 14F and the left side camera 14SL in the overhead view image, the vehicle speed of the vehicle V, and the rotation angle (steering angle) of the steering wheel 28. Then, a predicted rear wheel trajectory TR is calculated.
In step S302, for example, a process of detecting parallelism between the straight line X and the length direction (for example, the depth direction) of the parking area L0 using an image captured by the front camera 14F among the overhead images (shown in the figure). “Parallelity detection”). If the process which detects the parallelism of the straight line X and the length direction of the parking area L0 is performed in step S302, the process which the parking area approach reliability setting part 38 performs will transfer to step S304.

Here, the parallelism detected in step S302 is detected as an angle θap formed by the center line Y and the straight line X of the parking area L0 as shown in FIG.
In step S302, when the vehicle V moves to the parking area L0 while moving backward, for example, the straight line X is parallel to the length direction of the parking area L0 using an image captured by the rear camera 14R in the overhead image. Process to detect the degree. Here, the moving direction (forward, backward) of the vehicle V is detected with reference to a current shift position signal, for example.
In step S304, processing for calculating the turning radius of the vehicle V using the vehicle speed of the vehicle V and the rotation angle (steering angle) of the steering wheel 28 ("turning radius calculation" shown in the figure) is performed. If the process which calculates the turning radius of the vehicle V is performed in step S304, the process which the parking area approach reliability setting part 38 performs will transfer to step S306.

In step S306, it is determined whether or not the parallelism (θap) detected in step S302 is less than a preset parallelism threshold (for example, 15 [°]) (“parallelism <parallel” shown in the figure). Degree threshold? ").
If it is determined in step S306 that the parallelism (θap) detected in step S302 is equal to or greater than the parallelism threshold (“No” in the drawing), the process performed by the parking area approach certainty setting unit 38 is step S308. Migrate to
On the other hand, when it is determined in step S306 that the parallelism (θap) detected in step S302 is less than the parallelism threshold (“Yes” shown in the drawing), the process performed by the parking area approach certainty setting unit 38 is: The process proceeds to step S310.

In step S308, it is determined whether or not the turning radius detected in step S304 is greater than or equal to a preset turning radius threshold (for example, 100 [R]) (“turning radius ≧ turning radius threshold? ")I do.
If it is determined in step S308 that the turning radius detected in step S304 is less than the turning radius threshold value ("No" shown in the figure), the processing performed by the parking area approach certainty setting unit 38 proceeds to step S312. .
On the other hand, if it is determined in step S308 that the turning radius detected in step S304 is equal to or greater than the turning radius threshold (“Yes” shown in the figure), the process performed by the parking area approach certainty setting unit 38 proceeds to step S310. Transition.

In step S310, a process for determining whether or not the amount of deviation detected in step S300 is greater than or equal to a preset first threshold (for example, 75 [cm]) (“deviation amount ≧ first threshold? ")I do.
If it is determined in step S310 that the amount of deviation detected in step S300 is greater than or equal to the first threshold ("Yes" shown in the figure), the process performed by the parking area entry certainty setting unit 38 proceeds to step S314. .
On the other hand, if it is determined in step S310 that the amount of deviation detected in step S300 is less than the first threshold ("No" shown in the figure), the process performed by the parking area entry certainty setting unit 38 proceeds to step S316. Transition.

In step S312, a process for determining whether or not the deviation amount detected in step S300 is greater than or equal to a preset second threshold (for example, 150 [cm]) (“deviation amount ≧ second threshold? ")I do. Here, the second threshold value is larger than the first threshold value described above.
In step S312, when it is determined that the amount of deviation detected in step S300 is equal to or greater than the second threshold ("Yes" shown in the figure), the process performed by the parking area entry certainty setting unit 38 proceeds to step S318. .
On the other hand, if it is determined in step S312 that the amount of deviation detected in step S300 is less than the second threshold ("No" shown in the figure), the process performed by the parking area entry certainty setting unit 38 proceeds to step S314. Transition.

In step S314, a process of setting the parking area approach certainty factor to a low level ("entry certainty factor = level low" shown in the figure) is performed. If the process which sets a parking area approach reliability to a low level is performed in step S314, the process which the parking area approach reliability setting part 38 performs will be complete | finished (END).
In step S316, a process of setting the parking area approach certainty factor to a high level ("entry certainty factor = level high" shown in the figure) is performed. If the process which sets parking area approach reliability to a high level is performed in step S316, the process which the parking area approach reliability setting part 38 performs will be complete | finished (END).

In step S318, a process of setting the parking area approach certainty level to the lowest value (level 0) (“entry certainty = level 0” shown in the figure) is performed. If the process which sets a parking area approach reliability to level 0 is performed in step S318, the process which the parking area approach reliability setting part 38 performs will be complete | finished (END).
As described above, the parking area approach certainty setting unit 38 sets the parking area approach certainty of the minimum value “level 0”, the level “level low” higher than level 0, and the level higher than level low. A process of setting one of the “high levels” is performed.

Process Performed by Total Confidence Setting Unit 40 The process in which the comprehensive certainty setting unit 40 sets the total certainty will be described using FIG. 30 with reference to FIGS. 1 to 29.
The overall certainty setting unit 40 receives the parking area certainty signal and the parking area entering certainty signal, and receives the parking area certainty included in the parking area certainty signal and the parking area entering certainty included in the parking area entering certainty signal. The degree is adapted to the comprehensive confidence setting map shown in FIG. Then, based on the parking area certainty and the parking area approach certainty, the overall certainty is set.
In FIG. 30, the parking area certainty is indicated as “frame certainty”, and the parking area entry certainty is indicated as “entry certainty”. 30 is a map used when the vehicle V travels forward.
As an example of the process in which the comprehensive certainty setting unit 40 sets the comprehensive certainty, the case where the parking area certainty is “level 3” and the parking area approach certainty is “level high” is shown in FIG. In this way, the overall certainty factor is set to “high”.

In the present embodiment, as an example, when the overall confidence setting unit 40 performs a process of setting the overall confidence, the set overall confidence is stored in a storage unit in which data is not erased even when the ignition switch is turned off. A case where the storing process is performed will be described. Here, the storage unit from which data is not erased even when the ignition switch is turned off is, for example, a nonvolatile memory such as a flash memory.
Therefore, in this embodiment, when the ignition switch is turned off after the parking of the vehicle V is completed and the ignition switch is turned on when the vehicle V restarts, the total certainty factor set immediately before is stored. For this reason, it is possible to start control based on the overall certainty factor set immediately before the ignition switch is turned on when the vehicle V restarts.

-Process which acceleration suppression control start timing calculating part 42 performs The process which the acceleration suppression control start timing calculating part 42 calculates an acceleration suppression control start timing is demonstrated using FIG. 31, referring FIGS. 1-30.
The acceleration suppression control start timing calculation unit 42 receives an input of the comprehensive certainty factor signal, and adapts the comprehensive certainty factor included in the comprehensive certainty factor signal to the acceleration suppression condition calculation map shown in FIG. Then, the acceleration suppression control start timing is calculated based on the total certainty factor.
In FIG. 31, the acceleration suppression control start timing is indicated as “suppression control start timing (accelerator opening)” in the “acceleration suppression condition” column.
As an example of the processing performed by the acceleration suppression control start timing calculation unit 42, when the overall certainty factor is “high”, the acceleration suppression control start timing is increased by increasing the opening of the accelerator pedal 32 as shown in FIG. Then, the timing is set to reach “50%”. The opening degree of the accelerator pedal 32 is set to 100% when the accelerator pedal 32 is depressed (operated) to the maximum value.

Processing performed by the acceleration suppression control amount calculation unit 44 With reference to FIGS. 1 to 31, processing in which the acceleration suppression control amount calculation unit 44 calculates the acceleration suppression control amount will be described.
The acceleration suppression control amount calculation unit 44 receives the input of the total certainty factor signal, and adapts the total certainty factor included in the total certainty factor signal to the acceleration suppression condition calculation map shown in FIG. Then, an acceleration suppression control amount is calculated based on the total certainty factor. In FIG. 31, the acceleration suppression control amount is indicated as “suppression amount” in the “acceleration suppression condition” column.

As an example of the processing performed by the acceleration suppression control amount calculation unit 44, when the total certainty factor is “high”, as shown in FIG. 31, the acceleration suppression control amount is set as the actual opening degree of the accelerator pedal 32. Set to a control amount that reduces the throttle opening to a “medium” level. In the present embodiment, as an example, the “medium” level throttle opening is set to 25% of the actual opening of the accelerator pedal 32. Similarly, the “small” level throttle opening is set to 50% of the actual accelerator pedal 32 opening, and the “large” level throttle opening is set to 10% of the actual accelerator pedal 32 opening. Of the opening.
Further, the acceleration suppression control amount calculation unit 44 sets the presence / absence of control to output a warning sound by adapting the total certainty factor to the acceleration suppression condition calculation map. In the case of outputting a warning sound, for example, character information on the content that activates the acceleration suppression control and visual information such as a symbol and light emission may be displayed on a display monitor included in the navigation device 26.

(Processing performed by the acceleration suppression command value calculation unit 10J)
Next, processing performed by the acceleration suppression command value calculation unit 10J will be described with reference to FIGS. 1 to 31 and FIG.
The acceleration suppression command value calculation unit 10J performs the processing described below for each preset sampling time (for example, 10 [msec]).
As shown in FIG. 32, when the acceleration suppression command value calculation unit 10J starts processing (START), first, in step S400, an acceleration suppression operation condition determination result signal received from the acceleration suppression control content calculation unit 10I is displayed. refer. And the process ("acceleration suppression operation condition judgment result acquisition process" shown in the figure) which acquires an acceleration suppression operation condition judgment result is performed. If the process which acquires an acceleration suppression operation condition judgment result is performed in step S400, the process which the acceleration suppression command value calculating part 10J performs will transfer to step S402.

In step S402, in addition to the acceleration suppression operation condition determination result acquired in step S400, processing for acquiring information for calculating the acceleration suppression command value ("acceleration suppression command value calculation information acquisition processing" shown in the figure) is performed. . If the process which acquires the information for calculating an acceleration suppression command value in step S402 is performed, the process which the acceleration suppression command value calculating part 10J performs will transfer to step S404.
The information for calculating the acceleration suppression command value is information included in the above-described acceleration suppression control start timing signal, acceleration suppression control amount signal, driving side depression amount signal, and accelerator operation speed signal, for example.

In step S404, a process of determining whether or not the acceleration suppression operation condition determination result acquired in step S400 is a determination result that the acceleration suppression control operation condition is satisfied (“acceleration suppression control operation condition satisfied?” Shown in the figure). Do.
If it is determined in step S404 that the acceleration suppression control operation condition is satisfied ("Yes" shown in the drawing), the processing performed by the acceleration suppression command value calculation unit 10J proceeds to step S406.
On the other hand, if it is determined in step S404 that the acceleration suppression control operation condition is not satisfied ("No" shown in the figure), the processing performed by the acceleration suppression command value calculation unit 10J proceeds to step S408.

In step S406, based on the information for calculating the acceleration suppression command value acquired in step S402, a process of calculating an acceleration suppression command value that is an acceleration command value for performing acceleration suppression control ("Acceleration suppression command shown in the figure"). Control command value calculation "). If the process which calculates an acceleration suppression command value is performed in step S406, the process which the acceleration suppression command value calculating part 10J performs will transfer to step S410.
Here, in the process of calculating the acceleration suppression command value, the depression amount of the accelerator pedal 32 included in the drive side depression amount signal and the acceleration suppression control amount included in the acceleration suppression control amount signal are referred to. Then, an acceleration suppression control amount command value is calculated that sets the throttle opening to a degree of suppression (see FIG. 31) corresponding to the acceleration suppression control amount with respect to the actual opening of the accelerator pedal 32.

Further, in the process of calculating the acceleration suppression command value, the depression amount of the accelerator pedal 32 included in the driving side depression amount signal and the acceleration suppression control start timing included in the acceleration suppression control start timing signal are referred to. And the acceleration suppression control start timing command value which makes acceleration suppression control start timing the timing (refer FIG. 31) according to the opening degree of the actual accelerator pedal 32 is calculated.
In the process of calculating the acceleration suppression command value, the command value including the acceleration suppression control amount command value and the acceleration suppression control start timing command value calculated as described above is calculated as the acceleration suppression command value.

In step S408, driving force control without acceleration suppression control, that is, processing for calculating a normal acceleration command value that is an acceleration command value used in normal acceleration control ("command value calculation for normal acceleration control" shown in the figure). I do. If the process which calculates a normal acceleration command value is performed in step S408, the process which the acceleration suppression command value calculating part 10J performs will transfer to step S412.
Here, in the process of calculating the normal acceleration command value, the command value for calculating the throttle opening based on the depression amount of the accelerator pedal 32 included in the drive side depression amount signal is calculated as the normal acceleration command value.

In step S410, an acceleration suppression command value signal including the acceleration suppression command value calculated in step S406 is output to the target throttle opening calculation unit 10K ("acceleration suppression command value output" shown in the figure). If the process which outputs an acceleration suppression command value signal is performed in step S410, the process which the acceleration suppression command value calculating part 10J performs will be complete | finished (END).
In step S412, a process of outputting a normal acceleration command value signal including the normal acceleration command value calculated in step S408 to the target throttle opening calculation unit 10K ("normal acceleration command value output" shown in the figure) is performed. If the process which outputs a normal acceleration command value signal is performed in step S412, the process which the acceleration suppression command value calculating part 10J performs will be complete | finished (END).

(Processing performed by the target throttle opening calculation unit 10K)
Next, processing performed by the target throttle opening calculation unit 10K will be described using FIG. 33 with reference to FIGS.
The target throttle opening calculation unit 10K performs the processing described below for each preset sampling time (for example, 10 [msec]).
As shown in FIG. 33, when the target throttle opening calculation unit 10K starts processing (START), first, in step S500, the drive side depression amount signal received from the accelerator operation amount calculation unit 10G is referred to. And the process ("accelerator operation amount acquisition process" shown in a figure) which acquires the depression amount (operation amount) of the accelerator pedal 32 which the drive side depression amount signal contains is performed. If the process which acquires the depression amount (operation amount) of the accelerator pedal 32 is performed in step S500, the process which the target throttle opening calculating part 10K performs will transfer to step S502.

In step S502, based on the information signal received from the acceleration suppression command value calculation unit 10J, the suppression with acceleration suppression command value (see step 408), the acceleration suppression command value (see step S414) or the normal acceleration command value (step S418). (Refer to) is performed ("command value acquisition process" shown in the figure). If the process which acquires an acceleration suppression command value or a normal acceleration command value is performed in step S502, the process which the target throttle opening calculating part 10K performs will transfer to step S504.
In step S504, calculation of the target throttle opening ("target throttle opening calculation" shown in the figure) is performed based on the depression amount of the accelerator pedal 32 acquired in step S500 and the command value acquired in step S502. When the target throttle opening is calculated in step S504, the processing performed by the target throttle opening calculation unit 10K proceeds to step S506.

Here, in step S504, when the command value acquired in step S502 is a normal acceleration command value (when the acceleration suppression operation condition is not established), the throttle opening corresponding to the depression amount of the accelerator pedal 32 is set as follows. Calculated as the target throttle opening.
On the other hand, when the command value acquired in step S502 is the acceleration suppression command value (when the acceleration suppression operation condition is satisfied), the throttle opening corresponding to the acceleration suppression control amount command value is set as the target throttle opening. Calculate.
The target throttle opening is calculated using, for example, the following equation (1).
θ * = θ1−Δθ (1)
In the above equation (1), the target throttle opening is indicated by “θ *”, the throttle opening corresponding to the depression amount of the accelerator pedal 32 is indicated by “θ1”, and the acceleration suppression control amount is indicated by “Δθ”.

In step S506, a target throttle opening signal including the target throttle opening θ * calculated in step S504 is output to the engine controller 12 (“target throttle opening output” shown in the figure). In step S506, when the process of outputting the target throttle opening signal to the engine controller 12 is performed, the process performed by the target throttle opening calculation unit 10K ends (END).
Here, in step S506, when the command value acquired in step S502 is an acceleration suppression command value, the opening (depression amount) of the accelerator pedal 32 reaches the opening corresponding to the acceleration suppression control start timing. The target throttle opening signal is output.

(Operation)
Next, an example of an operation performed using the vehicle acceleration suppression device 1 of the present embodiment will be described with reference to FIGS. 1 to 33 and FIG.
First, an example in which the vehicle V traveling in the parking lot enters the parking area selected by the driver will be described. Here, the selected parking area is a parking area composed of L-shaped ends shown in FIG. 34 (a) (in the following description, it may be referred to as "L-shaped end parking area"). Suppose that
In the state where the vehicle speed of the vehicle V traveling in the parking lot is 15 [km / h] or more which is the threshold vehicle speed, the acceleration suppression control operation condition is not satisfied. Normal acceleration control reflecting the person's intention to accelerate.

  On the other hand, when the vehicle speed is less than the threshold vehicle speed and the vehicle V travels toward the L-shaped end parking area, a part of the L-shaped end parking area is included in the overhead image captured by the surrounding environment recognition sensor 14. become. Here, as shown in FIG. 34 (a), it is assumed that the end portion Pmf indicated by a dotted line in the drawing has disappeared due to aged deterioration due to tire rubbing or the like. Accordingly, the bird's-eye view image includes only one L-shaped end portion Pnf closer to the vehicle V. The parking area certainty setting unit 36 performs an end determination process on the bird's-eye view image including the end Pnf (step S2000). Thereby, the edge part Pnf is first extracted as an edge part candidate. However, in this case, since the end candidate pair cannot be extracted, the parking area certainty setting unit 36 sets the end determination flag to OFF.

On the other hand, the parking area certainty degree setting unit 36 sets the area line determination flag to OFF in the area line determination process (step S2010) because no parking area line candidate is extracted from the overhead image.
Next, the parking area certainty degree setting unit 36 performs a first numeral determination process (step S2020).
That is, the parking area certainty setting unit 36 first acquires end part extraction information (step S2100), and determines whether an end part candidate has been extracted (step S2110). Here, since one edge candidate Pnf has been extracted (Yes in step S2110), the parking area certainty setting unit 36 next performs a numeric area extraction process (step S2120).

Specifically, the parking area certainty setting unit 36 first has a distance from a virtual line orthogonal to the perspective direction through the lower end of the end candidate Pnf in the overhead image as shown in FIG. Edge detection is performed on the area Ar1 in ds2 using a lower luminance threshold than other areas (see FIG. 20A).
In this L-shaped end parking area, as shown in FIG. 34A, the parking area identification number “40” is marked in the area Ar1. Therefore, the parking area certainty setting unit 36 detects the edge of “40” from the overhead image. Subsequently, edge detection is performed with respect to a region up to a distance ds3 farther than the distance ds2 by using a luminance threshold value that increases as the distance increases (see FIG. 20A). Further, edge detection is performed on a region beyond the distance ds3 by using a luminance threshold value that is constant at the value of ds3 (see FIG. 20A). Here, it is assumed that the edge region other than the identification number “40” is not detected.

  When the edge detection ends, the parking area certainty setting unit 36 then performs pattern matching processing on the image area (edge aggregate image) corresponding to the detected edge area. At this time, since the detected edge group image is in the area Ar1, the parking area certainty setting unit 36 performs pattern matching on the edge group image using a lower matching threshold than other areas ( (Refer FIG.20 (b)). Thereby, the numerical area NA including the numeral “40” is detected, and the parking area certainty setting unit 36 extracts the numerical area NA from the overhead image.

Since the number area NA has been extracted (Yes in step S2130), the parking area certainty level setting unit 36 next determines whether the number “40” included in the number area NA satisfies the condition of the number used for the parking area. A process of determining whether or not is performed (step S2140).
Specifically, the aspect ratio of the number “40” is compared with a preset aspect ratio range. Here, since the identification number of the L-shaped end parking area is extracted, the aspect ratio of the number “40” is within the range of the aspect ratio. Thereby, the aspect ratio flag is set to ON. In addition, since the size of the number “40” is also equal to or smaller than the size threshold, the size flag is set to ON.

Accordingly, the parking area certainty setting unit 36 determines that the numbers included in the extracted number area NA satisfy the conditions of the numbers used for the parking area (Yes in step S2140).
Subsequently, the parking area certainty setting unit 36 determines whether or not the distance ds4 between the extracted number area NA and the end candidate Pnf is within the first distance range. Here, the numerical area NA corresponding to the extracted edge candidate Pnf is extracted, and the distance ds4 is within the first distance range (Yes in step S2150).
Subsequently, the parking area certainty setting unit 36 determines whether or not the deviation from the reference position between the orientation of the extracted numeric area NA and the orientation of the end candidate Pnf is equal to or less than a preset deviation threshold. As shown in FIG. 34A, since there is no deviation in the direction, the parking area certainty setting unit 36 determines that the deviation in both directions is equal to or less than the deviation threshold (Yes in step S2160).

Next, the parking area certainty setting unit 36 has not extracted an edge candidate pair (Yes in step S2170), and subsequently, there is a non-numeric area in the area Ar3 surrounding the upper and left ends of the numeric area. It is determined whether or not it exists (step S2180).
The parking area certainty setting unit 36 determines whether or not a non-numeric area exists based on the edge area (edge aggregate image) detected when the numeric area NA is extracted. Here, since an edge area other than the numeric area is not detected, the parking area certainty setting unit 36 determines that there is no non-numeric area (No in step S2180).
Thereby, the parking area certainty setting part 36 sets a 1st number determination flag to ON (step S2190).
Subsequently, the parking area certainty level setting unit 36 performs a second number determination process (step S2030).
Here, since parking area line candidates are not extracted (No in step S2210), the parking area certainty setting unit 36 sets the second number determination flag to OFF (step S2290).

Since the first number determination flag is currently ON, the parking area certainty level setting unit 36 sets the parking area certainty level from the initial level 0 to level 1 ("Yes" in step S206, S208). .
Thus, when the L-shaped end parking area is detected, the brake pedal 30 is not operated, and the amount of depression of the accelerator pedal 32 is equal to or greater than the threshold accelerator operation amount, the vehicle V is parked at the L-shaped end. Determine whether to enter the area.
In addition, during traveling of the vehicle V, in addition to the setting of the parking area certainty factor of the parking area certainty factor setting unit 36, the parking area approach certainty factor setting unit 38 sets the parking area approach certainty factor. And the comprehensive reliability setting part 40 sets the comprehensive reliability based on a parking area reliability and a parking area approach reliability.

Further, while the vehicle V is traveling, the acceleration suppression control start timing calculation unit 42 calculates the acceleration suppression control start timing based on the total reliability set by the total reliability setting unit 40, and the acceleration suppression control amount calculation unit 44. Calculates the acceleration suppression control amount.
Then, when it is determined that the vehicle V enters the L-shaped end parking area and it is determined that the acceleration suppression control operation condition is satisfied, the acceleration suppression command value calculation unit 10J sends the acceleration suppression command value signal to the target throttle opening calculation unit 10K. Output to. Further, the target throttle opening calculation unit 10K outputs a target throttle opening signal to the engine controller 12.
At this time, since the parking area certainty is level 1, the total certainty is “very low” as shown in FIG. Therefore, as shown in FIG. 31, the suppression control start timing is “80 [%]”, the suppression amount is “small”, and the warning sound is “none”.

That is, the acceleration suppression control start timing is set to a timing at which the opening of the accelerator pedal 32 increases and reaches “80 [%]”, and the throttle opening is set to 50 [% of the actual opening of the accelerator pedal 32. ] Opening degree, and setting that does not output warning sound.
For this reason, when the driver operates the accelerator pedal 32 in a state where the acceleration suppression control operation condition is satisfied, the acceleration suppression control amount command value is subtracted from the throttle opening corresponding to the depression amount of the accelerator pedal 32, and the throttle is opened. The degree is set to 50 [%] of the actual throttle opening. Thereby, the acceleration which generate | occur | produces in the vehicle V is reduced and the acceleration of the vehicle V is suppressed. In addition to this, the start timing for reducing the throttle opening corresponding to the amount of depression of the accelerator pedal 32 (suppressing acceleration) is set to the timing corresponding to the acceleration suppression control start timing command value (the throttle opening reaches 80 [%]). Timing).

Subsequently, when the travel distance of the vehicle V reaches the set travel distance in the state of continuous collation, the parking area certainty setting unit 36 sets the parking area certainty from level 1 to level 2 (“Yes in step S210) ”, S212).
Subsequently, it is assumed that the vehicle V travels toward the L-shaped end parking area, and two end portions Pmb and Pnb on the side far from the vehicle V are included in the overhead view image.
As a result, the edge candidate pairs Pmb and Pnb are extracted. Therefore, in the edge determination process, it is determined that the edge condition is met, and the edge determination flag is set to ON.

Therefore, at this time, it is determined by the edge determination process that the edge candidate pairs Pmb and Pnb on the side far from the vehicle V are extracted and meet the conditions of the edges constituting the parking area. Therefore, the parking area certainty setting unit 36 sets the parking area certainty from level 2 to level 3 (“Yes” in step S214, S216).
Subsequently, since the end portion Pmf on the left front side cannot be extracted, the facing condition of both end points is not met (No in step S218), and the parking area certainty remains at level 3.
Thereby, for example, when the approach certainty level is “high”, the total certainty factor is set to “high” as shown in FIG. 30. Therefore, the acceleration suppression control operation condition is satisfied, and as shown in FIG. 31, the suppression control start timing is “50 [%]”, the suppression amount is “medium”, and the warning sound is “present”.

That is, the acceleration suppression control start timing is set to a timing at which the opening of the accelerator pedal 32 increases and reaches “50 [%]”, and the throttle opening is set to 25 [% of the actual opening of the accelerator pedal 32. ] Is set to output a warning sound.
For this reason, when the driver operates the accelerator pedal 32 in a state where the acceleration suppression control operation condition is satisfied, the acceleration suppression control amount command value is subtracted from the throttle opening corresponding to the depression amount of the accelerator pedal 32, and the throttle is opened. The degree is set to 25 [%] of the actual throttle opening. Thereby, the acceleration which generate | occur | produces in the vehicle V is reduced and the acceleration of the vehicle V is suppressed. In addition to this, the start timing for reducing the throttle opening (acceleration is suppressed) according to the amount of depression of the accelerator pedal 32 (the throttle opening reaches 50 [%]) according to the acceleration suppression control start timing command value. Timing). Further, a warning sound such as a buzzer sound or a warning message is output from the speaker of the information presentation device.

Next, the driver selects and parks a parking area (which may be described as “two-line parking area” in the following description) composed of two straight lines as shown in FIG. The operation will be described.
When the vehicle speed becomes less than the threshold vehicle speed and the vehicle V travels toward the two-line parking area, a part of the two-line parking area is included in the overhead image captured by the surrounding environment recognition sensor 14. Here, as shown in FIG. 34 (b), it is assumed that the parking area line Lm on the left side indicated by a dotted line in the drawing has disappeared due to aged deterioration due to tire rubbing or the like. Therefore, only one parking area line Ln on the right side is included in the overhead view image.

The parking area certainty setting unit 36 first performs an edge determination process on the bird's-eye view image including the parking area line Ln (step S2000). In this case, since only the front side end portion of the parking region line candidate Ln is extracted as the end portion candidate, the parking region certainty setting unit 36 sets the end portion determination flag to OFF.
Subsequently, the parking area certainty level setting unit 36 performs an area line determination process (step S2010). Thereby, the parking area line Ln is extracted as a parking area line candidate. However, in this case, since the area line candidate pair cannot be extracted, the parking area certainty level setting unit 36 sets the area line determination flag to OFF.

Next, the parking area certainty degree setting unit 36 performs a first numeral determination process (step S2020).
Here, since one end candidate has been extracted (Yes in step S2110), the parking area certainty setting unit 36 performs the first number determination through the same processing as in the L-shaped end parking area. The flag is set to ON (step S2190).
Subsequently, the parking area certainty level setting unit 36 performs a second number determination process (step S2030).
That is, the parking area certainty setting unit 36 first acquires area line extraction information (step S2200), and determines whether or not a parking area line candidate has been extracted (step S2210). Here, since one parking area line candidate Ln is extracted (Yes in step S2210), the parking area certainty setting unit 36 next performs a numerical area extraction process (step S2220).

Specifically, the parking area certainty setting unit 36 first determines the distance from the virtual line orthogonal to the perspective direction through the lower end of the parking area line candidate Ln in the overhead image as shown in FIG. Edge detection is performed on the area Ar1 in ds2 using a lower luminance threshold than other areas (see FIG. 20A).
In the two-line parking area, as shown in FIG. 34 (b), the parking area identification number “40” is drawn in the area Ar1. Therefore, the parking area certainty setting unit 36 detects the edge of “40” from the overhead image. Subsequently, edge detection is performed with respect to a region farther than the distance ds2 by using a luminance threshold that increases as the distance increases (see FIG. 20A). Here, it is assumed that the edge region other than the identification number “40” is not detected.

  When the edge detection ends, the parking area certainty setting unit 36 then performs pattern matching processing on the image area (edge aggregate image) corresponding to the detected edge area. At this time, since the detected edge group image is in the area Ar1, the parking area certainty setting unit 36 performs pattern matching on the edge group image using a lower matching threshold than other areas ( (Refer FIG.20 (b)). Thereby, the numerical area NA including the numeral “40” is detected, and the parking area certainty setting unit 36 extracts the numerical area NA from the overhead image.

Since the number area NA has been extracted (Yes in step S2230), the parking area certainty setting unit 36 next determines whether or not the number “40” included in the number area NA satisfies the condition of the number used for the parking area. The process which determines is performed (step S2240).
Specifically, the aspect ratio of the number “40” is compared with a preset aspect ratio range. Here, since the identification number of the two-line parking area is extracted, the aspect ratio of the number “40” is within the range of the aspect ratio. Thereby, the aspect ratio flag is set to ON. In addition, since the size of the number “40” is also equal to or smaller than the size threshold, the size flag is set to ON.

Therefore, the parking area certainty setting unit 36 determines that the numbers included in the extracted number area NA satisfy the conditions for the numbers used in the parking area (Yes in step S2240).
Subsequently, the parking area certainty setting unit 36 determines whether or not the distance ds10 between the extracted number area NA and the parking area line candidate Ln is within the first distance range. Here, the numeric area NA corresponding to the extracted parking area line candidate Ln is extracted, and the distance ds10 is within the first distance range (Yes in step S2250).
Subsequently, the parking area certainty setting unit 36 determines whether or not the deviation from the basic position between the direction of the extracted numeric area NA and the direction of the parking area line candidate is equal to or less than a preset deviation threshold. As shown in FIG. 34B, since there is no deviation in the direction, the parking area certainty setting unit 36 determines that the deviation in both directions is equal to or less than the deviation threshold (Yes in step S2260).

Next, since the area line candidate pair has not been extracted (Yes in step S2270), the parking area certainty level setting unit 36 then continues to the non-numeric area in the area Ar3 surrounding the upper end and the left and right ends of the numeric area NA. Is determined (step S2280).
The parking area certainty setting unit 36 determines whether or not there is a non-numeric area based on the edge area detected when the numeric area NA is extracted. Here, since no edge area other than the numeric area NA is detected, the parking area certainty setting unit 36 determines that there is no non-numeric area (No in step S2180).
As a result, the parking area certainty level setting unit 36 sets the second number determination flag to ON (step S2290).

Since the first number determination flag and the second number determination flag are both ON at the present time, the parking area certainty level setting unit 36 sets the parking area certainty level from the initial level 0 to level 1 (step S206). “Yes”, S208).
When it is determined that the vehicle V enters the two-line parking area and the acceleration suppression control operation condition is satisfied, the acceleration suppression command value calculation unit 10J outputs an acceleration suppression command value signal to the target throttle opening calculation unit 10K. To do. Further, the target throttle opening calculation unit 10K outputs a target throttle opening signal to the engine controller 12.

At this time, since the parking area certainty is level 1, the total certainty is “very low” as shown in FIG. Therefore, as shown in FIG. 31, the suppression control start timing is “80 [%]”, the suppression amount is “small”, and the warning sound is “none”.
Subsequently, when the travel distance of the vehicle V reaches the set travel distance in the state of continuous collation, the parking area certainty setting unit 36 sets the parking area certainty from level 1 to level 2 (“Yes in step S210) ”, S212).
Thereafter, since the left parking area line Lm cannot be extracted, it does not meet the facing conditions of the far and far end points and the both end points (No in step S214, No in step S218), and the parking area reliability remains at level 2. Become.

Thereby, for example, when the approach certainty level is “high”, the total certainty factor is set to “low” as shown in FIG. Therefore, the acceleration suppression control operation condition is satisfied, and as shown in FIG. 31, the suppression control start timing is “80 [%]”, the suppression amount is “medium”, and the warning sound is “present”.
That is, the acceleration suppression control start timing is set to a timing when the opening of the accelerator pedal 32 increases and reaches “80 [%]”, and the throttle opening is set to 25 [% of the actual opening of the accelerator pedal 32. ] Is set to output a warning sound.

  For this reason, when the driver operates the accelerator pedal 32 in a state where the acceleration suppression control operation condition is satisfied, the acceleration suppression control amount command value is subtracted from the throttle opening corresponding to the depression amount of the accelerator pedal 32, and the throttle is opened. The degree is set to 25 [%] of the actual throttle opening. Thereby, the acceleration which generate | occur | produces in the vehicle V is reduced and the acceleration of the vehicle V is suppressed. In addition to this, the start timing for reducing the throttle opening corresponding to the amount of depression of the accelerator pedal 32 (suppressing acceleration) is set to the timing corresponding to the acceleration suppression control start timing command value (the throttle opening reaches 80 [%]). Timing). Further, a warning sound such as a buzzer sound or a warning message is output from the speaker of the information presentation device.

Next, as the number area NA, for example, an operation when an area including numbers included in the school zone time zone as shown in FIG. 21B is extracted will be described.
As shown in FIG. 34 (c), among the numbers indicating the school zone time zone “7:30”, the number “7” has disappeared due to deterioration over time, and only the portion of the number “30” is extracted as a number area. Suppose that
Here, in the road marking in the school zone, the number representing the “minute” of the time is drawn smaller than the number representing the “hour”. Therefore, the aspect ratio and size of the numbers may fall within the range of the aspect ratio and size used for the parking area. As a result, it is determined that the number “30” included in the extracted number area NA satisfies the condition of the number used in the parking area (No in step S2140).

Here, it is assumed that only one edge candidate is extracted, and it is determined that the distance ds10 between the number area NA and the edge candidate is within the first distance range (Yes in step S2150). In addition, it is assumed that the deviation from the reference position between the direction of the number area NA and the direction of the edge candidate is determined to be equal to or less than the deviation threshold (Yes in step S2160). Further, since only one edge candidate is extracted, it is determined that the ratio between the width between edge candidates and the width of the number area NA is equal to or greater than the ratio threshold value (Yes in step S2170).
Next, the parking area certainty setting unit 36 determines whether or not there is a non-numeric area in an area Ar3 that surrounds the upper and left ends of the extracted numeric area NA and extends in the depth direction.

Here, in the edge detection process when extracting the numerical area NA, other edge areas are detected in addition to the numerical area NA, and the other edge areas are non-numeric areas that do not include numbers by pattern matching. Suppose that it is determined that.
The parking area certainty degree setting unit 36 determines whether or not the detected non-numeric area is included in the area Ar3. Here, as shown in FIG. 34 (c), since a part of the characters “ZONE” of “SCHOOL ZONE” is included in the area Ar3, the parking area certainty setting unit 36 has a non-numeric area. It is determined that it exists (Yes in step S2180).
As a result, the parking area certainty level setting unit 36 sets the second number determination flag to OFF (step S2195).

As described above, in the present embodiment, the parking area line includes a parking area composed of a parking area line such as a two-line parking area and a parking area composed only of an end portion such as an L-shaped end parking area. Even if a part of or a part of the end part is faded or disappeared, the parking area can be detected from the identification number of the parking area and the remaining parking area line or end.
In addition, based on the parking area detected in this manner, the accelerator pedal 32 may be operated due to an erroneous operation or the like in a situation where the braking operation is an appropriate driving operation such as a state in which the vehicle V approaches a position suitable for parking in the parking area. Even if it is operated, it is possible to reduce the throttle opening according to the total certainty factor and to suppress the acceleration of the vehicle V. That is, since the acceleration suppression amount (the degree of throttle opening reduction) is small when the overall confidence level is low, it is possible to reduce the reduction in drivability, and when the overall confidence level is high, the acceleration suppression amount is large. Therefore, the acceleration suppression effect of the vehicle V can be increased.

In other words, in the present embodiment, during parking, it is possible to suppress a decrease in drivability in the parking lot before entering the parking area, and to suppress acceleration of the vehicle V when the accelerator pedal 32 is erroneously operated. Is possible.
In the present embodiment, the higher the total certainty factor, the greater the acceleration suppression control amount, thereby suppressing the acceleration of the vehicle V and improving the safety. Further, the lower the overall certainty, the later the acceleration suppression control start timing is delayed, and the drivability is suppressed from decreasing. This makes it possible to improve safety and suppress deterioration of drivability under the following conditions.

For example, in a situation where the vehicle V standing by in the vicinity where a parking area for parallel parking is marked on the side of the traveling road on the road, it is necessary to allow some acceleration.
Even under the following conditions, it is necessary to allow a certain amount of acceleration. This is because there are other vehicles on both sides (left and right parking areas) of the parking area where the vehicle V is parked, and the vehicle V enters from a front side into a small space on the opposite side (side away from each parking area). Thereafter, the vehicle V is parked by entering the parking area where the vehicle V is parked from the rear side.
By controlling the acceleration suppression control start timing and the acceleration suppression control amount based on the total certainty for these situations, it is possible to suppress acceleration of the vehicle V and improve safety. In addition, it is possible to allow acceleration of the vehicle V and suppress drivability deterioration.

  In addition, although control which reduces the acceleration command value which accelerates the vehicle V was mentioned as an example and demonstrated as acceleration suppression control, it is not restricted to this structure. For example, the acceleration suppression control includes control for causing the vehicle V to travel at a low vehicle speed equal to or lower than a preset vehicle speed, and control for decelerating (including stopping) the vehicle V by a braking device as well as driving force control. Further, the acceleration suppression control includes power transmission control based on clutch connection control (for example, when suppression is performed, the clutch is disconnected from the gear and power is not transmitted).

Here, the accelerator pedal 32 described above corresponds to an acceleration operator, and the amount of depression of the accelerator pedal 32 described above corresponds to an acceleration operation amount.
The accelerator operation detection sensor 24 and the accelerator operation amount calculation unit 10G described above correspond to an acceleration operation amount detection unit.
The acceleration suppression command value calculation unit 10J described above and the target throttle opening calculation unit 10K described above correspond to an acceleration control unit.
The ambient environment recognition sensor 14 described above corresponds to an imaging unit.
The end candidate extraction unit 36a described above corresponds to the end candidate extraction unit, the number area extraction unit 36c corresponds to the number area extraction unit, and the parking area detection unit 36d corresponds to the parking area detection unit. The parking area certainty factor calculation unit 36e corresponds to a parking area certainty factor calculation unit.
Moreover, the acceleration suppression control start timing calculation unit 42, the acceleration suppression control amount calculation unit 44, the acceleration suppression command value calculation unit 10J, and the target throttle opening calculation unit 10K described above correspond to the acceleration suppression control unit.
Further, the template image described above corresponds to a matching image, and the matching threshold value described above corresponds to a matching degree threshold value.

(Effect of embodiment)
If it is this embodiment, it will become possible to show the effect described below.
(1) The accelerator operation detection sensor 24 and the accelerator operation amount calculation unit 10G detect the depression amount (acceleration operation amount) of the driver's accelerator pedal 32. The acceleration suppression command value calculation unit 10J and the target throttle opening calculation unit 10K control the acceleration generated in the vehicle V according to the depression amount of the accelerator pedal 32. The surrounding environment recognition sensor 14 images a region including the road surface around the vehicle V. The parking end candidate extraction unit 36a extracts end candidates that are candidates for the ends constituting the parking area from the captured image captured by the surrounding environment recognition sensor 14. The number area extraction unit 36c extracts a number area NA including a symbol string including numbers from the captured image. The parking area detection unit 36d determines that the distance ds4 (and ds5) between the extracted edge candidate and the numeric area NA is within the first distance range when the distance ds4 (and ds5) is within the preset first distance range. A parking area composed of the end candidate and the numerical area NA is detected (end determination flag is set to ON). When the acceleration suppression control start timing calculation unit 42, the acceleration suppression control amount calculation unit 44, the acceleration suppression command value calculation unit 10J, and the target throttle opening calculation unit 10K detect the parking area, The acceleration generated in the vehicle V is reduced (acceleration command value (throttle opening) is reduced) according to the acceleration operation amount (depression amount) detected by the accelerator operation detection sensor 24 and the accelerator operation amount calculation unit 10G.

With this configuration, the parking area end candidate and the numeric area NA located on the road surface are extracted from the captured image obtained by imaging the area including the road surface around the host vehicle, and between the extracted end candidate and the numeric area NA. It is possible to detect the parking area comprised of the end candidate and the numeric area NA whose distance ds4 (and ds5) is within the preset first distance range. And if a parking area | region is detected, it is possible to implement the control (acceleration suppression control) which reduces the acceleration generated in the vehicle V according to the operation amount of the accelerator pedal 32 of a driver | operator.
Here, as shown in FIGS. 6A and 6B, when detecting a parking area composed of only four parking area ends arranged at the four corners, as described in the end determination process. First, an image region (end candidate) that satisfies the above-described conditions B1 to B2 located on the road surface in the overhead image is extracted. At this time, if two adjacent edge candidates (edge candidate pairs) can be extracted from the bird's-eye view image within a preset interval range, parking can be performed by determining the shape, orientation, etc. of these edge candidate pairs. It is possible to detect the region with relatively high accuracy.

However, when only one edge candidate is extracted, even if the shape of the edge candidate is suitable as the parking area edge, there is a sign pattern that does not constitute a parking area of the same shape on the road surface. There is a high possibility that it is marked, and there is not enough judgment material to determine that it constitutes a parking area.
In addition, for example, a part of the parking area end part constituting the parking area may disappear or fade due to deterioration over time, and only one end candidate corresponding to the parking area end part may be extracted. is there.

In addition, the parking area formed on the structure whose surface becomes like a mirror surface when wet with water, such as concrete, gets wet by rain, etc., and the edge corresponding to the parking area edge by reflection on the surface of surrounding objects In some cases, candidates cannot be detected successfully.
Moreover, even if it is a parking area partitioned by the parking area line, a case where most of the parking area line disappears due to deterioration over time or the like, and only the end portion of the marking remains. For example, when the entire parking area line on one side of the parking area composed of two parallel parking area lines disappears, most of the remaining one disappears and only the end part remains. Etc. are considered. In such a case, the indicated end portion is extracted as an end candidate.

On the other hand, some parking areas are labeled with an identification number for identifying the parking area. This identification number is a unique number for each parking area, and is roughly similar in size, marking position, and the like.
In the case of the above configuration, for example, when a sign (number) identifying a parking area remains under bad conditions where only an edge candidate corresponding to one parking area edge can be extracted, the number It is possible to detect a parking area with higher accuracy than determining from only one end candidate by determining the distance relationship between the number area including the number and the extracted at least one end candidate. .
And when the vehicle V is advancing with respect to such a parking area, when the erroneous operation of an accelerator generate | occur | produces, it becomes possible to reduce the acceleration of the vehicle V. FIG. As a result, it is possible to suppress the occurrence of acceleration not intended by the driver when the vehicle V is parked in the parking area.

(2) The parking area detection unit 36d, from the captured image, the end of a straight line (parking area line candidate) that matches the conditions (the above conditions C1 to C3) as the preset parking area line, and the preset parking An image pattern that conforms to conditions (the above-mentioned conditions B1 to B2 and the above-described preset shape pattern) as a marking pattern of a predetermined shape that is arranged at the corner of the area and divides the parking area is extracted as an edge candidate.
With this configuration, in addition to the end of the parking area line that constitutes the parking area, for example, for a parking area that is composed only of an L-shaped or U-shaped pattern, some of these patterns disappear, Even when it is hazy, the parking area can be detected from the numeric area including the identification number and the extracted sign pattern.

(3) The deviation from the preset reference angle between the direction of the symbol string including the number of the extracted number area NA and the direction of the extracted end candidate is equal to or less than a preset deviation threshold. And the parking area | region comprised from the edge part candidate and the number area | region NA whose distance ds10 is in the 1st distance range is detected.
With this configuration, it is determined whether or not the direction of the symbol string including numbers and the direction of the extracted end candidate are the relationship between the numbers (identification numbers) constituting the parking area and the parking area end. Is possible.
As a result, the detection accuracy of the parking area can be further improved, and it is possible to more appropriately suppress the occurrence of acceleration not intended by the driver when the vehicle V is parked in the parking area.

(4) The parking area detection unit 36d determines whether or not the aspect ratio of the numbers included in the extracted number area is within a preset aspect ratio, and parks the number area that is determined not to be within the range. Exclude from area detection.
Here, on the public road, there are numbers such as speed limit, school zone time zone, bus exclusive road time zone, U-turn time zone, and the like. Among these numbers, for example, a number indicating the speed limit is marked vertically in comparison with a number used as a parking area identification number.
Based on this, for example, a range of an aspect ratio not including an aspect ratio of a number that is not normally used as a number indicating an identification number of a parking area such as an aspect ratio of a number indicating a speed limit is set in advance. Whether or not the aspect ratio of the numbers included in the number area NA is included in the range is determined. Then, the numeric area NA determined not to be included in the range of the aspect ratio is excluded from the parking area detection target.
As a result, the detection accuracy of the parking area can be further improved, and it is possible to more appropriately suppress the occurrence of acceleration not intended by the driver when the vehicle V is parked in the parking area.

(5) The parking area detection unit 36d determines whether or not there is a non-numeric area that is an area including a symbol other than a numeral within a preset second distance range from the extracted numeric area NA in the captured image. . If the parking area detection unit 36d determines that a non-numeric area exists, the extracted numeric area NA is excluded from the parking area detection targets.
Here, the time zone of the school zone and the time zone of the bus exclusive road marked on the public road are marked with a character string of “school zone” or a character string of “bus exclusive” around the time zone number. Has been.
Based on this, it is determined whether or not there is a non-numeric area that includes a symbol other than a numeral within the second distance range from the numeric area NA. Excluded from area detection.
As a result, the detection accuracy of the parking area can be further improved, and it is possible to more appropriately suppress the occurrence of acceleration not intended by the driver when the vehicle V is parked in the parking area.

(6) The number area extraction unit 36c extracts the number area NA based on pixels in the captured image that are equal to or higher than a preset luminance threshold. The number area extraction unit 36c sets the luminance threshold to a lower value as the pixel has a shorter distance from the extracted edge candidate.
Here, the identification number of the parking area is often marked near the end on the near side (for example, on the side opposite to the wheel stopper (parking block)) of the ends constituting the parking area.
Based on this, the luminance threshold value is lowered as the pixel has a shorter distance from the front end. Thereby, it becomes possible to easily extract candidates (edge set images) of the numerical area NA.
That is, for example, it is possible to improve the extraction accuracy of the numerical area NA including the identification number that has become faint due to aging or the like. As a result, the detection accuracy of the parking area can be further improved, and it is possible to more appropriately suppress the occurrence of acceleration not intended by the driver when the vehicle V is parked in the parking area.

(7) The number area extraction unit 36c selects an image area whose matching degree (matching rate) with a matching image (template image) set in advance from the captured image is equal to or higher than a preset matching degree threshold (matching threshold). Extract as NA. The number area extraction unit 36c sets the matching degree threshold (matching threshold) to a lower value as the image area has a shorter distance from the extracted edge candidate.
Here, the identification number of the parking area is often marked near the end on the near side (for example, on the side opposite to the wheel stopper (parking block)) of the ends constituting the parking area.
Based on this, the pattern matching matching threshold is set lower for image regions that are candidates for the number region existing at a position where the distance from the end on the near side is short. This makes it easier to extract the numeric area NA.
That is, for example, it is possible to improve the extraction accuracy of the numerical area NA including the identification number that has become faint due to aging or the like. As a result, the detection accuracy of the parking area can be further improved, and it is possible to more appropriately suppress the occurrence of acceleration not intended by the driver when the vehicle V is parked in the parking area.

(8) The parking area detection unit 36d determines whether or not the size of the symbol string including the numbers in the extracted number area NA is equal to or larger than a predetermined size threshold, and determines that it is equal to or larger than the size threshold. The numeric area NA is excluded from the parking area detection targets.
Here, the numbers marked on public roads are often larger than the numbers used as identification numbers for parking areas.
Based on this, the number area NA is compared with a preset size threshold, and the number area NA including a number that is not used as the size of the identification number of the parking area is parked. Excluded from area detection. As a result, the detection accuracy of the parking area can be further improved, and the occurrence of acceleration not intended by the driver when the vehicle V is parked in the parking area can be more appropriately suppressed.

(9) The ratio between the width of the extracted numeric area NA and the width between the end candidate pairs facing each other across the numeric area NA among the extracted end candidates is preset by the parking area detection unit 36d. It is determined whether or not the ratio is equal to or greater than the threshold value. In addition, the numerical area NA determined that the ratio between the width between the end candidate pairs and the extracted numerical area NA is equal to or larger than the ratio threshold is excluded from the parking area detection targets.
Here, the ratio between the width of the identification number of the parking area and the width of the two opposite ends constituting the parking area is the width between the number marked on the public road and the end of the road white line marked on the public road. The ratio with the width is different. Specifically, the ratio corresponding to the parking area is smaller than the ratio corresponding to the public road.
Based on this, if it is determined that the ratio between the width of the extracted number area NA and the width between the extracted end candidate pairs is equal to or greater than a preset ratio threshold, the number area NA is detected from the detection target of the parking area. Excluded. As a result, the detection accuracy of the parking area can be further improved, and the occurrence of acceleration not intended by the driver when the vehicle V is parked in the parking area can be more appropriately suppressed.

(10) The parking area certainty degree calculation unit 36e calculates a parking area certainty degree indicating the degree of certainty that the parking area exists in the traveling direction of the host vehicle based on the detected parking area. The acceleration suppression control start timing calculation unit 42, the acceleration suppression control amount calculation unit 44, the acceleration suppression command value calculation unit 10J, and the target throttle opening calculation unit 10K are calculated by the parking region certainty degree setting unit 36. Based on the certainty factor, when the parking region certainty factor is low, the degree of reduction in acceleration is made smaller than when the parking region certainty factor is high. That is, when the parking area certainty degree calculated by the parking area certainty degree setting unit 36 is high, the degree of reduction of the acceleration command value is increased compared to when the parking area certainty degree is low.
With this configuration, it is possible to set the parking area certainty based on the detection result of the parking area having the identification number. In addition, in a state where the set parking area certainty factor is low, it is possible to reduce the decrease in the acceleration command value to reduce the drivability, and in a state where the set parking area certainty factor is high, the acceleration command value It is possible to increase the acceleration suppression effect of the vehicle V by increasing the degree of reduction of the vehicle V.
As a result, it is possible to more appropriately suppress the occurrence of acceleration not intended by the driver when the vehicle V is parked in the parking area.

(11) In the vehicle acceleration suppression method according to the present embodiment, from a captured image obtained by capturing an area around the vehicle V, an end candidate that is a candidate for an end constituting the parking area, and a number including a symbol string including numbers The area NA is extracted. In addition, when the distance ds4 (and ds5) between the extracted edge candidate and the number area NA is within the preset first distance range, the edge candidate determined to be within the first distance range; A parking area composed of a numerical area NA is detected. When the parking area is detected, the acceleration generated in the vehicle V is reduced according to the acceleration operation amount (depression amount) of the driver's acceleration operator (accelerator pedal 32).
Accordingly, it is possible to detect a parking area having an identification number based on the extracted at least one end candidate and the numeric area NA. And it becomes possible to implement acceleration suppression control at the time of parking the vehicle V in such a parking area. As a result, it is possible to suppress the occurrence of acceleration not intended by the driver when the vehicle V is parked in the parking area.

(Modification)
(1) In the above-described embodiment, the first number determination process is performed using the information on the end candidate extracted in the end determination process. However, the present invention is not limited to this configuration. For example, the first number determination process may be configured to perform a process of extracting edge candidates from the overhead image separately from the edge determination process.
(2) In the above embodiment, the second number determination process is performed using the information of the parking area line candidates extracted in the area line determination process, but the present invention is not limited to this configuration. For example, it is good also as a structure which implements the process which extracts a parking area line candidate from an overhead image separately from an area line determination process in a 2nd number determination process.

(3) In the above embodiment, the edge threshold or the parking area line candidate is extracted first, and the luminance threshold used for edge detection is determined according to the distance from the virtual line orthogonal to the perspective direction through the near side end. Although the configuration is such that the lower the distance is, the lower the configuration is, the configuration is not limited thereto. For example, the luminance threshold value may not be changed according to the distance from the virtual line. Further, in the case of this configuration, the same luminance threshold value may be used to simultaneously detect an edge that is an edge candidate or a parking area line candidate and an edge that is a number area candidate. Or it is good also as a structure which performs the process which extracts a number area | region first, and extracts an edge part candidate or a parking area line candidate continuously, for example, when a number area | region is extracted. In addition, the luminance threshold value is changed. First, the edge candidate or the parking area line candidate and the edge as the number area candidate are detected using the same luminance threshold value. After that, for example, when an edge that becomes a numeric area is not detected, or when an edge that seems to be a numeric area is slightly detected, from an imaginary line that passes through the front end of the extracted edge candidate or parking area line candidate Alternatively, the brightness threshold value may be changed based on the distance and the edge detection process may be performed again.

(4) In the above embodiment, the edge threshold or the parking area line candidate is extracted first, and the luminance threshold value used for edge detection is determined according to the distance from the virtual line orthogonal to the perspective direction through the near side end. The matching threshold used for pattern matching is changed. For example, a configuration in which only one of them is changed, such as a configuration in which only the matching threshold value is changed, may be used.
(5) In the above embodiment, when there is a non-numeric area within the second distance range from the numeric area NA, the numeric area NA is excluded from the detection target of the parking area. However, the present invention is not limited to this configuration. . For example, even when there is a symbol string including a number within the second distance range from the numerical area NA, the numerical area NA may be excluded from detection targets. That is, it is based on the fact that the time zone of the bus exclusive road or school zone is marked in two steps in the depth direction, but the parking area numbers are rarely marked in two steps.

(6) In the above embodiment, in the perspective direction through the line extending in the perspective direction through the left and right center of the extracted number area NA and the left and right center of the extracted end candidate or the left and right center of the extracted parking area line candidate. It is determined whether or not the distance to the extending line is within the first distance range. Thus, in the said embodiment, although it was set as the structure which determines the distance of the left-right direction between number area | region NA and an edge part candidate or a parking area line candidate, it is not restricted to this structure. For example, the distance in the perspective direction is set in advance in addition to the distance in the left-right direction based on the fact that the identification number of the parking area is often present in the vicinity of the front end of the end candidate or the parking area line candidate. It is good also as a structure which determines whether it is in the range. In this case, for example, a virtual line when changing the luminance threshold value or the matching threshold value according to the distance is set, and it is determined whether the distance from the virtual line is within the distance range.

(7) In the above embodiment, for example, as shown in FIGS. 18A to 18D, various processing contents have been described by taking the case where the identification number is marked inside the parking area as an example. The configuration is not limited to this. For example, as shown to FIG.18 (e)-(h), it can be set as the structure which implements the same process also with respect to the structure by which the identification number is labeled on the outer side of a parking area. In this case, when determining the distance between the numerical area NA and the edge candidate, for example, as shown by the solid double-pointed arrows in FIGS. 35 (a) and 35 (b), It is determined whether the distances ds4 and ds5 between the line extending in the perspective direction through the point and the line extending in the perspective direction through the left and right center points of the end candidates Pmf and Pnf are within the first distance range. Also, taking into account the distance in the perspective direction, as shown by the dotted double arrow lines in FIGS. 35 (a) and (b), the left and right midpoints of the front end of the end candidates Pmf and Pnf and the numerical area NA It is determined whether or not the linear distances ds4 and ds5 between the center point are within the first distance range. When determining the distance between the numeric area NA and the parking area line candidate, for example, as shown by the solid double-pointed arrows in FIGS. 35 (c) and 35 (d), the midpoint in the horizontal direction of the numeric area NA. It is determined whether the distances ds10 and ds11 between the line extending in the perspective direction through the line and the line extending in the perspective direction through the left and right center points of the parking area line candidates Lm and Ln are within the first distance range. Also, taking into account the distance in the perspective direction, as shown by the dotted double arrow lines in FIGS. 35C and 35D, the left and right midpoints at the front side ends of the parking area line candidates Lm and Ln and the numerical area NA It is determined whether or not the linear distances ds10 and ds11 between the center points of the two are within the first distance range.

(8) In the above embodiment, if any one of the edge determination flag, the area line determination flag, the first numeral determination flag, and the second numeral determination flag is in the ON state, the parking area certainty is set to “level”. The configuration is set from “0” to “level 1”, but is not limited to this configuration. For example, when both the edge determination flag and the first numeral determination flag are in the ON state, or when both the area line determination flag and the second numeral determination flag are in the ON state, the parking area certainty is set to “level 0. The level may be set to a higher level based on a combination of flags, such as setting from “” to “level 2”.

(9) In the above embodiment, the number area NA is extracted and then the number size included in the number area NA is determined. However, the present invention is not limited to this configuration. For example, a size threshold value (for example, area) of the number area to be extracted is set in advance, and an image area portion including a number or a symbol string including a number that is equal to or larger than the set size threshold value is excluded from the number area candidates. It is good also as composition to do.
(10) In the above embodiment, the acceleration suppression control start timing and the acceleration suppression control amount are calculated based on the total reliability set by the total reliability setting unit 40. However, the present invention is not limited to this. That is, the acceleration suppression control start timing and the acceleration suppression control amount may be calculated based only on the parking area reliability set by the parking area reliability setting unit 36. In this case, the acceleration suppression control start timing and the acceleration suppression control amount are calculated by adapting the parking area certainty to, for example, an acceleration suppression condition calculation map shown in FIG.

(11) In the above embodiment, the configuration of the parking area certainty setting unit 36 is configured to set the parking area certainty based on the bird's-eye view image (environment) around the vehicle V and the vehicle speed (running state) of the vehicle V. However, the configuration of the parking area certainty setting unit 36 is not limited to this. That is, the configuration of the parking area certainty setting unit 36 is added to the bird's-eye view around the vehicle V and the vehicle speed, and the vehicle V including the current position of the vehicle V included in the own vehicle position signal and the vehicle road information signal is traveling. It is good also as a structure which sets parking area reliability using the classification (road classification) of the road to perform.
In this case, for example, when it is detected that the current position of the vehicle V is on a public road based on information included in the own vehicle position signal and the traveling road information signal, it is determined that there is no parking area L0 around the vehicle V, The parking area certainty is set to “level 0”.
Thereby, for example, when the vehicle V enters a parking area where the operation of the acceleration suppression control is not preferable, such as a parking area arranged on the road edge on a public road, it is possible to suppress the drivability reduction of the vehicle V. Become.

(12) In the above embodiment, if the parking area certainty setting unit 36 determines that the ends are opposed to each other along the direction of the width WL with respect to the lines La and Lb, the parking area certainty is determined. Processing to set to level 3 or level 4 is performed. However, the process of setting the parking area certainty level to level 3 or level 4 is not limited to this. That is, when the end shape of the line L is, for example, a U-shape (see FIGS. 5 (g) to (k), (m), (n)), the shape is not marked on the public road. When it is recognized that there is, the parking area certainty may be set to level 3 or level 4. The same applies to a parking area (see FIG. 6B) that includes only end portions.

(13) In the above embodiment, the parking area certainty factor setting unit 36 is configured to set the parking region certainty factor based on the bird's-eye view image (environment) around the vehicle V and the vehicle speed (running state) of the vehicle V. However, the configuration of the parking area certainty setting unit 36 is not limited to this. That is, for example, when the configuration of the vehicle V includes a device (parking support device) that assists the driver in steering to the parking region L0, if the parking support device is in the ON state, the parking region It is good also as a structure which becomes easy to raise the level of certainty. Here, the configuration in which the level of the parking area certainty is likely to increase is, for example, a configuration in which the above-described set movement distance is set to a shorter distance than usual.

(14) In the above embodiment, the acceleration suppression control amount and the acceleration suppression control start timing are changed based on the total certainty factor, and the reduction degree of the acceleration command value is changed. However, the present invention is not limited to this. That is, only the acceleration suppression control start timing or only the acceleration suppression control amount may be changed according to the total certainty factor, and the degree of reduction of the acceleration command value may be changed. In this case, for example, the higher the total certainty factor, the larger the acceleration suppression control amount may be set, and the acceleration command value reduction degree may be increased without changing the acceleration suppression control start timing.
(15) In the above embodiment, the acceleration command value is controlled to suppress the driving force of the vehicle V in accordance with the amount of depression of the accelerator pedal 32 (acceleration operation amount). However, the present invention is not limited to this. That is, for example, the throttle opening corresponding to the depression amount (acceleration operation amount) of the accelerator pedal 32 is set as the target throttle opening, and further, the braking force is generated by the braking device described above, and the vehicle according to the driving force operation amount. The driving force of V may be suppressed.

(16) In the above embodiment, the parking area certainty factor is calculated as a level 0 that is the lowest value and a level (levels 1 to 4) that is higher than the lowest value. However, the present invention is not limited to this. That is, the parking area certainty factor may be calculated as only two levels: a level that is the lowest value (for example, “level 0”) and a level that is higher than the lowest value (for example, “level 100”).
(17) In the above embodiment, the parking area approach certainty factor is calculated as “level 0” as the lowest value, “level low” as a level higher than level 0, and “level high” as a level higher than level low. The parking area entry certainty level is not limited to this. That is, the parking area approach certainty may be calculated as only two levels: a level that is the lowest value (for example, “level 0”) and a level that is higher than the lowest value (for example, “level 100”).

(18) In the above-described embodiment, the total confidence factor is divided into four stages according to the parking area confidence factor calculated as one of five levels and the parking area approach confidence factor calculated as one of three levels. Level (“very low”, “low”, “high”, “very high”). However, the overall confidence level is not limited to this. In other words, the total certainty factor may be calculated as only two levels of a level that is the lowest value (for example, “level 0”) and a level that is higher than the lowest value (for example, “level 100”).
In this case, for example, when the parking area certainty and the parking area approach certainty are calculated as the lowest level, the total certainty is calculated as the lowest level. For example, if the parking area certainty and the parking area approach certainty are calculated as a level higher than the lowest value, the total certainty is calculated as a level higher than the lowest value.

(19) In the above embodiment, the parking area is detected based on the end candidate pair, the area line candidate pair, the parking area line candidate and the number area NA, and the end candidate and the number area NA. Although it was set as the structure, it is not restricted to this structure. For example, as shown in FIGS. 18 (a) to 18 (d), some parking areas are provided with vehicle stops (parking blocks) PB. Since the position of the vehicle stop PB is roughly determined, the parking area may be detected based on the positional relationship between the vehicle stop PB and the numerical area NA.

  The above embodiments are preferable specific examples of the present invention, and various technically preferable limitations are given. However, the scope of the present invention is described in particular in the above description to limit the present invention. As long as there is no, it is not restricted to these forms. In the drawings used in the above description, for convenience of illustration, the vertical and horizontal scales of members or parts are schematic views different from actual ones. In addition, the present invention is not limited to the above-described embodiments, and modifications, improvements, equivalents, and the like within the scope that can achieve the object of the present invention are included in the present invention.

DESCRIPTION OF SYMBOLS 1 Vehicle acceleration suppression apparatus 2 Brake apparatus 4 Fluid pressure circuit 6 Brake controller 8 Engine 10 Running control controller 10A Ambient environment recognition information calculation part 10B Own vehicle vehicle speed calculation part 10C Steering angle calculation part 10D Steering angular speed calculation part 10E Shift position calculation part 10F Brake pedal operation information calculation unit 10G Accelerator operation amount calculation unit 10H Acceleration operation speed calculation unit 10I Acceleration suppression control content calculation unit 10J Acceleration suppression command value calculation unit 10K Target throttle opening calculation unit 12 Engine controller 14 Ambient environment recognition sensor (front Camera 14F, right side camera 14SR, left side camera 14SL, rear camera 14R)
DESCRIPTION OF SYMBOLS 16 Wheel speed sensor 18 Steering angle sensor 20 Shift position sensor 22 Brake operation detection sensor 24 Accelerator operation detection sensor 26 Navigation apparatus 28 Steering wheel 30 Brake pedal 32 Accelerator pedal 34 Acceleration suppression operation condition judgment part 36 Parking area reliability setting part 36a End Part candidate extraction unit 36b Line extraction unit 36c Number region extraction unit 36d Parking region detection unit 36e Parking region certainty factor calculation unit 38 Parking region approach certainty factor setting unit 40 Total certainty factor setting unit 42 Acceleration suppression control start timing calculation unit 44 Acceleration suppression Control amount calculation unit V Own vehicle W Wheel (right front wheel WFR, left front wheel WFL, right rear wheel WRR, left rear wheel WRL)

Claims (11)

An acceleration operator that the driver operates to instruct acceleration;
An acceleration operation amount detector for detecting an acceleration operation amount of the acceleration operator;
An acceleration control unit that causes the host vehicle to generate acceleration according to the acceleration operation amount detected by the acceleration operation amount detection unit;
An imaging unit for imaging an area including a road surface around the host vehicle;
An edge candidate extraction unit that extracts edge candidates that are edge candidates constituting a parking area from a captured image captured by the imaging unit;
A number area extraction unit that extracts a number area including a symbol string including numbers from the captured image;
If the distance between the edge candidate extracted by the edge candidate extraction unit and the number area extracted by the number area extraction unit is within a preset first distance range, the distance is within the first distance range. A parking area detection unit for detecting a parking area composed of the end candidate and the number area;
An acceleration suppression control device for a vehicle, comprising: an acceleration suppression control unit that reduces the acceleration generated by the acceleration control unit when the parking region detection unit determines that the parking region has been detected.   The end candidate extraction unit is arranged from the captured image and arranged at the end of a line that meets a condition as a line constituting a preset parking area and at a corner of the preset parking area. The vehicle acceleration suppression device according to claim 1, wherein an image region that meets a condition as a sign pattern having a predetermined shape is extracted as the end candidate.   The parking area detection unit has a preset reference angle between the direction of the symbol string including the numbers in the number area extracted by the number area extraction unit and the direction of the end candidates extracted by the end candidate extraction unit And when the distance between the extracted edge candidate and the number area is within the first distance range, the deviation is less than the deviation threshold and the first 3. The vehicle acceleration suppression device according to claim 1, wherein a parking area configured by the end candidate and the number area within one distance range is detected.   The parking area detection unit determines whether or not the aspect ratio of the numbers included in the number area extracted by the number area extraction unit is within a preset aspect ratio range, and is determined not to be within the range. The vehicular acceleration suppression device according to any one of claims 1 to 3, wherein the area is excluded from the detection target of the parking area.   The parking area detection unit determines whether or not there is a non-numeric area that includes a symbol other than a numeral within a second distance range set in advance from the numeric area extracted in the captured image. The acceleration suppression device for a vehicle according to any one of claims 1 to 4, wherein a numeric area determined to have a numeric area is excluded from a detection target of a parking area.   The number area extraction unit is configured to extract the number area based on pixels that are equal to or higher than a preset brightness threshold in the captured image, and the distance from the edge candidate extracted by the edge candidate extraction unit is 6. The vehicle acceleration suppression device according to claim 1, wherein the shorter the pixel, the lower the luminance threshold value is set.   The number area extraction unit is configured to extract, as the number area, an image area in which the degree of coincidence with a preset matching image from the captured image is equal to or greater than a preset coincidence threshold. The vehicular acceleration suppression device according to any one of claims 1 to 6, wherein an image region having a shorter distance from the edge candidate extracted by the extraction unit sets the coincidence threshold value to a lower value. .   The parking area detection unit determines whether or not the size of the symbol string including the numbers included in the number area extracted by the number area extraction unit is greater than or equal to a predetermined size threshold, and is greater than or equal to the size threshold The acceleration suppression device for a vehicle according to any one of claims 1 to 7, wherein the numeric area determined to be is excluded from a detection target of a parking area.   The parking area detection unit includes a width of the numeric area extracted by the numeric area extraction unit and an end candidate extracted by the end candidate extraction unit between two end candidates facing each other across the numeric area. It is determined whether or not the ratio with the width is equal to or greater than a preset ratio threshold, and the number area for which the ratio is determined to be equal to or greater than the ratio threshold is excluded from the detection target of the parking area. Item 9. The vehicle acceleration suppression device according to Item 7 or 8. Based on the parking area detected by the parking area detection unit, a parking area reliability calculation unit that calculates a parking area reliability indicating the degree of certainty that the parking area exists in the traveling direction of the host vehicle,
The acceleration suppression control unit, based on the parking region certainty factor calculated by the parking region certainty factor calculating unit, reduces the acceleration reduction degree when the parking region certainty factor is low compared to when the parking region certainty factor is high. The acceleration suppression device for a vehicle according to any one of claims 1 to 9, wherein the acceleration suppression device is reduced.   Extracted end candidates from the captured image obtained by capturing the area including the road surface around the host vehicle are extracted as end candidates that are candidates for the end of the parking area and number areas including symbol strings including numbers. When the distance between the number area and the number area is within the preset first distance range, the parking area is detected by detecting the parking area composed of the end candidate and the number area within the first distance range. Then, the acceleration suppression method for vehicles characterized by reducing the acceleration of the own vehicle according to the acceleration operation amount of a driver | operator's acceleration operation element.

Images