JP2010170255A - Lane marking detecting device, marking detecting program, and lane departure warning device - Google Patents

Abstract

<P>PROBLEM TO BE SOLVED: To provide a lane marking detecting device which detects a lane marking with high accuracy, from a captured image, which is a borderline between cruising lanes where a vehicle travels. <P>SOLUTION: When detecting a lane marking, a lane departure warning device calculates likelihood which represents possibility that each lane marking candidate is a lane marking for each lane marking candidate representing a plurality of edge components arranged along the travelling direction of a vehicle from a captured image, based on a shape of each lane marking candidate (S320). Then, the lane departure warning device changes the likelihood corresponding to a lane marking candidate which is located near a preceding vehicle to a higher value (S330, S340), and identifies a lane marking candidate corresponding to the highest likelihood as a lane marking (S380). Accordingly, a lane marking is detected with high accuracy because a lane marking is detected by considering the location of the preceding vehicle. <P>COPYRIGHT: (C)2010,JPO&INPIT

Description

  The present invention relates to a lane boundary line detection device that detects a lane boundary line, a boundary line detection program, and a departure warning device that issues a warning when there is a possibility that a vehicle will depart from a traveling lane.

  A technique for detecting a lane boundary line that is a boundary of a traveling lane in which a vehicle travels from a captured image and generating an alarm when the vehicle is likely to deviate from the traveling lane is known (see, for example, Patent Document 1). ).

Japanese Patent No. 3332500

  By the way, in the above technique, paint on a road such as a white line is detected as a lane boundary from a captured image. However, on the actual road, there are various types of paint other than the paint indicating the lane boundary. For this reason, in the above technique, there is a possibility that a paint that does not indicate a lane boundary line is erroneously detected as a lane boundary line. When such a false detection occurs, an alarm is issued regardless of the departure from the lane, which is troublesome for the vehicle occupant.

  Accordingly, in view of such a problem, a lane boundary line detection device that detects a lane boundary line that is a boundary of a traveling lane in which a vehicle travels from a captured image, so that the lane boundary line can be accurately detected. It is an object of the present invention.

  The lane boundary line detection device according to claim 1, wherein the expected value calculation means is a traveling direction that represents a plurality of edge components arranged along the traveling direction of the vehicle in the captured image. For each component, a lane expectation value indicating the high possibility that each traveling direction component is a lane boundary line is calculated based on the shape of each traveling direction component. The expectation changing means changes the expected lane value corresponding to the traveling direction component in the vicinity of the position of the preceding vehicle representing the vehicle traveling ahead of the vehicle to a higher value, and the lane boundary line specifying means is The traveling direction component corresponding to the highest expected lane value is specified as the lane boundary line.

  That is, it can be estimated that the preceding vehicle is usually traveling in the traveling lane, so the traveling direction component near the position of the preceding vehicle is compared with the traveling direction component far away from the position of the preceding vehicle. There is a high possibility of a boundary line. Therefore, in the present invention, the traveling direction component and the position of the preceding vehicle are detected, and the expected lane value in the traveling direction component near the position of the preceding vehicle is changed so as to increase the possibility of being a lane boundary line.

  Therefore, according to such a lane boundary line detection apparatus, the lane boundary line is detected in consideration of the position of the preceding vehicle, so that the lane boundary line can be detected with high accuracy.

  Note that when the region as the traveling lane is identified based on the lane boundary line, the region on the center side of the lane boundary line identified in the captured image may be identified as the traveling lane.

  By the way, in the lane boundary line detection device according to claim 1, the expectation value changing means includes a plurality of traveling direction components having the same distance from the left or right end portion of the preceding vehicle as described in claim 2. When the vehicle is present inside and outside the preceding vehicle, the lane expected value of the traveling direction component located outside the preceding vehicle is compared with the amount of change in the expected lane value for the traveling direction component located inside the preceding vehicle. The change amount may be changed to a relatively high value.

  That is, in many cases, the preceding vehicle does not travel across the traveling lane. Therefore, in the present invention, the traveling direction component (the preceding vehicle straddles the inner side of the preceding vehicle from the left and right ends of the preceding vehicle). A higher lane expectation value is set for a traveling direction component (traveling direction component that the preceding vehicle does not straddle) located outside the traveling direction component).

  According to such a lane boundary line detection device, a higher lane expectation value is set for the traveling direction component that the preceding vehicle does not straddle, so in many cases where the preceding vehicle is traveling inside the traveling lane. The lane boundary line can be detected with higher accuracy.

  Furthermore, in the lane boundary line detection apparatus according to claim 1 or 2, the expected value calculation means sets the expected lane value to a higher value as the length of the traveling direction component is longer as described in claim 3. You may make it do.

  According to such a lane boundary detection device, the expected lane value is set to a small value for a sign or a character having a relatively short traveling direction component in the captured image, and the lengths of the traveling direction components are compared. For long white lines, etc., the expected lane value can be set to a large value. Therefore, it is possible to prevent erroneous detection of a sign, a character, or the like as a lane boundary line.

  Further, in the lane boundary line detection device according to any one of claims 1 to 3, the lane boundary line specifying unit, as described in claim 4, may be configured to capture the image on the basis of a position corresponding to the front of the vehicle. In each region obtained by dividing the region in the image into left and right, the traveling direction component corresponding to the highest expected lane value may be specified as the lane boundary line.

  According to such a lane boundary line detection device, since the lane boundary line is detected in each area obtained by dividing the area in the captured image into the left and right sides, when the lane boundary line is on both sides, each lane boundary line is detected. Can be detected.

  Next, a boundary line detection program according to a fifth aspect of the present invention causes a computer to function as each means constituting the lane boundary line detection device according to any one of the first to fourth aspects.

  According to such a boundary line detection program, since it has a function as each means which comprises the lane boundary line detection apparatus in any one of Claims 1-4, it is the lane of Claim 1 at least. The same effect as that of the boundary line detection device can be enjoyed.

  Next, in the departure warning device according to claim 5, wherein the lane boundary detection unit is configured to achieve the above-described object, the lane boundary detection means travels along a lane boundary that is a boundary of a traveling lane on which the vehicle travels. The direction is detected from the picked-up image. The departure determination means detects an approaching speed at which the vehicle approaches the lane boundary line, and determines a departure possibility indicating a possibility of departing from the detected traveling lane based on the approaching speed. Then, when the possibility of departure is higher than a predetermined reference value, the warning instruction means instructs the warning means for warning the passenger of the vehicle to issue a warning. Particularly in the present invention, the lane boundary line detection means is configured as the lane boundary line detection device according to any one of claims 1 to 5.

  According to such a departure warning device, since the lane boundary line detection device according to any one of claims 1 to 5 is provided, the lane boundary line can be detected with high accuracy. Therefore, it is possible to prevent the alarm that is performed based on the detection of the lane boundary line from malfunctioning.

1 is a block diagram (a) showing a schematic configuration of a traveling lane departure detection system to which the present invention is applied, and a block diagram (b) showing a configuration of a departure detection ECU. It is a flowchart which shows deviation warning processing. It is a flowchart which shows the lane boundary line candidate extraction process in the departure warning process. It is a flowchart which shows the lane boundary line selection process in the departure warning process. It is the map (a) which shows the relationship between the number of votes by Hough transform, and the probability (PFi), and the map (b) which shows the relationship between edge part distance and the probability (PPi). It is a flowchart which shows the deviation determination process among the deviation warning processes.

  Embodiments according to the present invention will be described below with reference to the drawings.

[First Embodiment]
[Configuration of this embodiment]
1A is a block diagram showing a schematic configuration of a traveling lane departure detection system 1 (deviation warning device) to which the present invention is applied, and FIG. 1B is a configuration of a departure detection ECU 10 (lane boundary line detection device). FIG. The traveling lane departure detection system 1 is mounted on a vehicle such as a passenger car, for example, and detects that the vehicle departs from the traveling lane. Note that the travel lane departure detection system 1 of the present embodiment also has a function of issuing an alarm when there is a possibility that the vehicle departs from the travel lane.

  More specifically, as shown in FIG. 1A, the lane departure detection system 1 includes a departure detection ECU (ECU is an electronic control unit) 10, a vehicle speed sensor 20, a navigation ECU 30, a vehicle control ECU 40 ( Alarm means) is connected to a communication line 5 on which communication is performed by a protocol such as CAN (Controller Area Network), for example, and is configured to be able to communicate with each other. A camera 15 and a radar 16 are connected to the departure detection ECU 10.

  The camera 15 is arranged so as to be able to image the road surface ahead of the vehicle. Further, the radar 16 transmits electromagnetic waves such as radio waves and laser light toward the front of the vehicle, and detects the position and schematic shape of the front object existing in front of the vehicle by detecting the reflected wave. It is configured as a well-known radar. The departure detection ECU 10 is configured to be able to acquire the image captured by the camera 15 and the radar 16 and the detection result of the front object at an arbitrary timing.

  The navigation ECU 30 is connected to an ETC (Electronic Toll Collection) vehicle-mounted device 35. The navigation ECU 30 receives information indicating that the ETC vehicle-mounted device 35 has passed the toll gate and toll gate passing information such as a fee settlement result. It is configured so that it can be acquired. Further, the navigation ECU 30 can provide toll gate passage information in response to a request from the departure detection ECU 10.

  As shown in FIG. 1B, the deviation detection ECU 10 is configured as a well-known microcomputer including a CPU 11, a ROM 12, a RAM 13, and the like, and the lane and the vehicle are connected via a communication interface (I / F) 14. A process of acquiring information necessary for processing to detect a lane boundary that is a boundary of a traveling lane on which the vehicle travels from another ECU or sensor, and detecting that the vehicle deviates from the traveling lane, and there is a risk of deviating Implement a process to issue an alarm. These processes are performed based on a program (including a boundary detection program) stored in the ROM 12.

  Note that the hardware configuration of the navigation ECU 30 and the vehicle control ECU 40 is the same as that of the departure detection ECU 10, and thus the description thereof is omitted. The vehicle control ECU 40 is connected to an alarm unit 45 configured as a speaker, an indicator lamp, or the like, and the vehicle control ECU 40 operates the alarm unit 45 in response to an instruction from the departure detection ECU 10 to drive the vehicle. It functions as an alarm device that warns the vehicle occupants of lane departures. However, the vehicle control ECU 40 can also function as a device for preventing a traffic accident such as a steering torque control device that controls the steering torque so that the vehicle does not deviate from the travel lane.

[Processing in this embodiment]
A process performed in the travel lane departure detection system 1 configured as described above and that issues a warning when the vehicle may deviate from the travel lane will be described with reference to FIG. 2 and subsequent drawings.

  2 is a flowchart showing a departure warning process executed by the departure detection ECU 10, FIG. 3 is a flowchart showing a lane boundary line candidate extraction process in the departure warning process, and FIG. 4 is a lane boundary selection process in the departure warning process. It is a flowchart to show. FIG. 6 is a flowchart showing a departure determination process in the departure warning process.

  The departure warning process is a process that is started when the power source of the vehicle such as an ignition switch is turned on, and then repeatedly executed at regular intervals (for example, every 100 ms). Specifically, as shown in FIG. 2, first, a lane boundary line that is a boundary (horizontal edge) of a traveling lane in which the vehicle travels is detected from a captured image in which the traveling direction of the vehicle is captured. (S110-S140: Lane boundary detection means). Specifically, first, a captured image in which the traveling direction of the vehicle is captured by the camera 15 is acquired from the camera 15 (S110).

  Then, a lane boundary line candidate extraction process for detecting a lane boundary line candidate (travel direction component) representing a plurality of edge components arranged along the traveling direction of the vehicle from a captured image obtained by capturing the traveling direction of the vehicle. Implement (S120: travel direction component detection means). In the lane boundary line candidate extraction process, as shown in FIG. 3, first, a differential value is calculated using a differential filter for each horizontal line (all pixels having the same vertical coordinate value) in the captured image (S210). ).

  That is, the change rate of the luminance value of an adjacent pixel is calculated for a plurality of pixels constituting the horizontal line. When the camera 15 is a color camera, the RGB signal output from the color camera or the change rate of the color difference signal when the RGB signal is converted into a luminance signal and a color difference signal may be calculated. .

  Subsequently, it is determined whether or not the calculated differential value is equal to or greater than a predetermined threshold value. If the differential value is equal to or greater than the threshold value, it is assumed that the luminance value of the adjacent pixel has greatly changed, and the coordinate value of the corresponding pixel is determined as an edge. A point (edge component) is registered in the RAM 13 (S220). After performing the processing of S210 and S220 for all the pixels in the captured image, the process proceeds to S230.

  Subsequently, a known Hough conversion process is performed on the basis of the registered edge points, so that the vehicle travels in the traveling direction (the angle (yaw angle) formed with the traveling direction of the vehicle is less than a certain value. )) Detect all linear patterns. Then, each line pattern is labeled as a lane boundary line candidate (S230).

  Labeling indicates a process of numbering each lane boundary line candidate such that i = 0, 1,... N. Subsequently, for each straight line pattern, the number of votes in the Hough transform process (that is, the total number of edge points used when detecting one straight line pattern) and the yaw angle information are recorded in the RAM 13 for each straight line pattern (S240). Then, the lane boundary line candidate extraction process ends.

  Next, returning to FIG. 2, the preceding vehicle detection process is performed (S130: preceding vehicle detection means). The preceding vehicle detection process is a process of detecting a preceding vehicle that travels ahead of the vehicle ahead of the vehicle, along with its position (a relative position with respect to the position of the lane boundary line candidate). Specifically, when detecting the preceding vehicle, a method using the detection result of the radar 16 or shape data recorded in advance in the ROM 12 or the like as the shape of the set of edge components in the captured image and the shape of the vehicle. Or a method of detecting a pair of taillights of a preceding vehicle from a captured image in the case of nighttime.

  Subsequently, a lane boundary line selection process for selecting a lane boundary line from a plurality of straight line patterns that are lane boundary line candidates is performed (S140).

  In the lane boundary line selection process, as shown in FIG. 4, first, one of the lane boundary line candidates extracted in the lane boundary line candidate extraction process (i = 0) is selected (S310). Next, the first probability (PFi, lane expected value) that the selected lane boundary line candidate is a lane boundary line is calculated based on the shape of the lane boundary line candidate (S320: expected value calculation means).

  In the present embodiment, the first probability is set to a higher value as the number of votes by the Hough transform recorded in the process of S240 increases. Specifically, as shown in FIG. 5A, when the number of votes by the Hough transform is determined, the first probability is calculated using a map in which the probability is uniquely determined. In the map shown in FIG. 5A, when the number of votes is 50 or less, the first probability is set to 0, and when the number of votes is 150 or more, the first probability is set to 0.5. It is set.

  When the number of votes is 50 or more and less than 150, the first probability is set to increase proportionally as the number of votes increases. In addition, it shows that it is a long straight line, so that the number of votes by Hough transformation increases.

  Subsequently, the second probability (PPi) based on the position of the preceding vehicle is calculated (S330). In the present embodiment, using the position of the lane boundary line candidate extracted in S120 and the position of the preceding vehicle detected in S130, either the left or right end of the lane boundary line candidate and the preceding vehicle The distance (hereinafter referred to as “end portion distance”) is detected using a map whose probability is uniquely determined according to this end portion distance. The probability (PPi) of 2 is calculated. In this map, when a plurality of lane boundary line candidates having the same end distance are present inside and outside the preceding vehicle, the second probability is compared with the lane boundary line candidate located inside the preceding vehicle. Thus, the second probability of the lane boundary line candidate located outside the preceding vehicle is set to a higher value.

  Specifically, as shown in FIG. 5B, the end distance is less than −0.25 m (when the lane boundary line candidate is located on the inner side (center side) of the preceding vehicle than the end portion of the preceding vehicle, When the end distance is a negative value) and the end distance is 1.2 m or more, the second probability (PPi) is set to 0 which is the minimum value, and the end distance is 0.25 to 0.25. If within the range of 1.0 m, the second probability is set to 0.2 which is the maximum value. If the end distance is within the range of -0.25 to 0.25 m, the second probability value is set proportionally larger as the end distance increases, and the end distance is 1.0 to In the range of 1.2 m, the value of the second probability is set proportionally smaller as the end distance increases.

  Subsequently, a total likelihood (Pi) that is the sum of the respective likelihoods (PFi and PPi) is calculated (S340). That is, in the processes of S330 and S340, the probability is further increased by adding the second probability (PPi) to the first probability (PFi) corresponding to the lane boundary line candidate in the vicinity of the position of the preceding vehicle. A process of changing to a higher value is performed (expected value changing means).

  Then, the calculated total probability value (Pi) is recorded in the RAM 13 (S350), and it is determined whether or not all the lane boundary line candidates have been selected (S360). If no lane boundary line candidate has been selected (S360: NO), the next lane boundary line candidate (i having a larger value of i by 1) is selected (S370), and the processes from S320 onward are repeated.

  On the other hand, if all the lane boundary line candidates have been selected (S360: YES), the lane boundary line candidate having the highest overall probability (Pi) is specified as the lane boundary line (S380: Lane boundary line) Specific means). At this time, the traveling direction component corresponding to the highest probability (Pi) in each of the regions divided into the left and right with reference to the position corresponding to the front of the vehicle in the captured image (for example, the image center in the left and right direction). Are identified as lane boundary lines.

  Then, the selected lane boundary line is recorded in the RAM 13 (S390), and the lane boundary line selection process is terminated.

  Subsequently, returning to FIG. 2, departure determination processing is performed (S150: departure determination means). This departure determination process is a process for detecting the approach speed at which the vehicle approaches the lane boundary and determining the possibility of departure indicating the possibility of deviating from the detected travel lane based on the approach speed. Specifically, as shown in FIG. 6, first, the approach speed at which the vehicle approaches the lane boundary is defined as the departure speed (Vd), and the departure speed (Vd) is set as the yaw angle (the extending direction of the lane boundary line). (The angle formed by the vehicle traveling direction) and the vehicle speed (S510). The yaw angle is obtained during the Hough conversion process in S230, and this value is recorded in the RAM 13 during the process in S240.

  Subsequently, a threshold value (dtlb) of the end distance (DTLB) is calculated (S520). This threshold value (dtlb) can be determined by the product of the departure speed (Vd) and the reference time (α) (for example, 2 seconds).

  Next, by comparing the end distance (DTLB) with the threshold value (dtlb), it is determined whether or not the vehicle departs from the traveling lane within the reference time (α) (S530).

  If the end distance (DTLB) is less than or equal to the threshold value (dtlb) (S530: YES), the departure possibility flag is set to the ON state in the RAM 13 as the vehicle departs from the traveling lane (S540), and the departure determination is made. The process ends. On the other hand, if the end portion distance (DTLB) is larger than the threshold value (dtlb) (S530: NO), it is assumed that the vehicle does not depart from the traveling lane, and the departure possibility flag is set to the OFF state in the RAM 13 (S550). The departure determination process is terminated.

  Subsequently, returning to FIG. 2, the departure possibility flag in the RAM 13 is referred to. When the departure possibility flag is in the ON state (when the departure possibility is higher than a predetermined reference value), The vehicle control ECU 40 is instructed to warn of the possibility of departure (S160: alarm instruction means). Upon receiving this instruction, the vehicle control ECU 40 immediately issues an alarm via the alarm unit 45.

  When such processing ends, the departure warning processing ends.

[Effects of this embodiment]
In the travel lane departure detection system 1 described in detail above, the departure detection ECU 10 captures the traveling direction of the vehicle in the lane boundary that is the boundary of the travel lane on which the vehicle travels in the departure warning process. Detect from the captured image. Further, the departure detection ECU 10 detects an approach speed at which the vehicle approaches the lane boundary line, and determines a possibility of departure indicating a possibility of deviating from the detected travel lane based on the approach speed. Then, when the possibility of departure is higher than a predetermined reference value, the vehicle control ECU 40 that is an alarm device that issues an alarm to an occupant of the vehicle is instructed to issue an alarm.

  In particular, when the departure detection ECU 10 detects a lane boundary line, in the departure warning process, first, for each lane boundary line candidate that represents a plurality of edge components arranged along the traveling direction of the vehicle in the captured image, A probability that indicates the likelihood that each lane boundary line candidate is a lane boundary line is calculated based on the shape of each lane boundary line candidate. Then, the probability corresponding to the lane boundary line candidate in the vicinity of the position of the preceding vehicle representing the vehicle traveling ahead of the vehicle is changed to a higher value, and the lane boundary line candidate corresponding to the highest probability is changed. , Identified as a lane boundary.

  That is, it can be assumed that the preceding vehicle is usually traveling in the traveling lane, so the lane boundary candidate near the position of the preceding vehicle is compared with the lane boundary candidate that is far away from the preceding vehicle position. This is likely to be a lane boundary. Therefore, in the present embodiment, the lane boundary line candidate and the position of the preceding vehicle are detected, and the probability of the lane boundary line candidate near the position of the preceding vehicle is changed so as to increase the possibility of being the lane boundary line. Yes.

  The vicinity of the position of the preceding vehicle takes into account the width of the preceding vehicle (about 1.6 m to 2.5 m) and the travel lane width (about 2 m to 4 m) in the width direction of the preceding vehicle (the left-right direction of the captured image). When the preceding vehicle is traveling in the traveling lane, the lane boundary line is estimated to exist.

  Therefore, according to such a traveling lane departure detection system 1, since the lane boundary line is detected in consideration of the position of the preceding vehicle, the lane boundary line can be detected with high accuracy. In addition, since the lane boundary line can be detected with high accuracy, it is possible to prevent the alarm that is performed based on the detection of the lane boundary line from malfunctioning.

  Note that when the region as the traveling lane is identified based on the lane boundary line, the region on the center side of the lane boundary line identified in the captured image may be identified as the traveling lane.

  Further, the departure detection ECU 10 determines that the departure warning process is performed on the inside of the preceding vehicle when there are a plurality of lane boundary line candidates having the same distance from the left or right end of the preceding vehicle on the inside and outside of the preceding vehicle. The amount of change in the probability of the lane boundary line candidate located outside the preceding vehicle is set to a relatively high value compared to the amount of change in the probability for the lane boundary line candidate located.

  That is, in many cases, the preceding vehicle does not travel across the traveling lane. Therefore, in the present embodiment, the lane boundary line candidate (the preceding vehicle is located on the inner side of the preceding vehicle from the left and right ends of the preceding vehicle). A higher probability is set for a lane boundary line candidate (a lane boundary line candidate that the preceding vehicle does not straddle) located outside the lane boundary line candidate).

  According to such a traveling lane departure detection system 1, since a higher probability is set for a lane boundary line candidate that the preceding vehicle does not straddle, in many cases, the preceding vehicle travels inside the traveling lane. In addition, the lane boundary line can be detected with higher accuracy.

  Further, the departure detection ECU 10 sets the probability to a higher value as the length of the lane boundary line candidate is longer.

  According to such a traveling lane departure detection system 1, the probability of a lane boundary line candidate in the captured image is set to a small value for a sign or a character with a relatively short lane boundary line length, and the length of the lane boundary line candidate is set. For a relatively long white line or the like, the probability can be set to a large value. Therefore, it is possible to prevent erroneous detection of a sign, a character, or the like as a lane boundary line.

  Further, the departure detection ECU 10 uses the lane boundary line candidate corresponding to the highest probability as the lane boundary line in each of the areas obtained by dividing the area in the captured image into the left and right with reference to the position corresponding to the front of the vehicle. Identify.

  According to the traveling lane departure detection system 1 as described above, since the lane boundary line is detected in each region obtained by dividing the region in the captured image into the left and right, when the lane boundary line is on both sides, each lane boundary line is detected. Can be detected.

[Other Embodiments]
Embodiments of the present invention are not limited to the above-described embodiments, and can take various forms as long as they belong to the technical scope of the present invention.

  For example, in the above embodiment, the first probability (PFi) calculated based on the shape of the lane boundary line candidate is calculated based on the number of votes in the Hough transform. For example, a straight line is given priority over a curve (higher) The probability may be set), or the solid line may be prioritized over the broken line. Further, when the information on the width of the traveling lane can be acquired from the navigation ECU 30, an area where the lane boundary line is detected is estimated according to the width of the traveling lane, and is higher than the lane boundary line candidate existing in this area. The certainty may be set.

  DESCRIPTION OF SYMBOLS 1 ... Travel lane deviation detection system, 5 ... Communication line, 10 ... Deviation detection ECU, 11 ... CPU, 12 ... ROM, 13 ... RAM, 15 ... Camera, 16 ... Radar, 20 ... Vehicle speed sensor, 30 ... Navigation ECU, 35 ... ETC on-board unit, 40 ... vehicle control ECU, 45 ... alarm unit.

Claims (6)

A lane boundary detection device that is mounted on a vehicle and detects a lane boundary that is a boundary of a traveling lane on which the vehicle travels,
A traveling direction component detecting means for detecting a traveling direction component representing a plurality of edge components arranged along the traveling direction of the vehicle from a captured image in which the traveling direction of the vehicle is captured;
Preceding vehicle detection means for detecting a preceding vehicle representing a vehicle traveling ahead of the vehicle together with its position;
Expected value calculation means for calculating, for each detected traveling direction component, a lane expected value that indicates the likelihood that each traveling direction component is the lane boundary line based on the shape of each traveling direction component. When,
Expected value changing means for changing the expected lane value corresponding to the traveling direction component in the vicinity of the detected position of the preceding vehicle to a higher value;
Lane boundary line specifying means for specifying a traveling direction component corresponding to the highest expected lane value as the lane boundary line;
A lane boundary line detection apparatus comprising: In the lane boundary line detection apparatus according to claim 1,
The expected value changing means is configured such that when a plurality of travel direction components having the same distance from the left or right end portion of the preceding vehicle are present inside and outside the preceding vehicle, the traveling direction is located inside the preceding vehicle. Lane boundary detection characterized in that the amount of change in the expected lane value of the driving direction component located outside the preceding vehicle is changed to a relatively high value compared to the amount of change in the expected lane value for the component apparatus. In the lane boundary line detection apparatus according to claim 1 or 2,
The expected value calculation means sets the expected lane value to a higher value as the length of the traveling direction component is longer. In the lane boundary line detection apparatus according to any one of claims 1 to 3,
The lane boundary line specifying unit is configured to obtain a traveling direction component corresponding to the highest expected lane value in each region obtained by dividing a region in the captured image into left and right with reference to a position corresponding to the front of the vehicle. A lane boundary line detection device characterized by being identified as a boundary line.   A boundary line detection program for causing a computer to function as each means constituting the lane boundary line detection device according to any one of claims 1 to 4. A departure warning device that is mounted on a vehicle and issues a warning when the vehicle may deviate from the driving lane,
Lane boundary detection means for detecting a lane boundary that is a boundary of a traveling lane in which the vehicle travels from a captured image in which the traveling direction of the vehicle is captured;
Deviation determination means for detecting an approaching speed at which the vehicle approaches the lane boundary line, and determining a departure possibility indicating a possibility of departing from the detected traveling lane based on the approaching speed;
An alarm instruction means for instructing an alarm means for alarming an occupant of the vehicle when the possibility of departure is higher than a predetermined reference value;
With
The lane boundary line detection means is configured as the lane boundary line detection device according to any one of claims 1 to 5.

Images