JP2009159241A - Parking-time stop assisting system - Google Patents

Abstract

<P>PROBLEM TO BE SOLVED: To provide a parking-time stop assisting system for stopping a vehicle securely in a set parking position in a parking lot which is frequently used although neither parking section lines nor vehicle stoppers are available. <P>SOLUTION: The parking-time stop assisting system includes: imaging means 1a and 1b of acquiring captured images of a vehicle periphery as suitable parking images when a vehicle is suitably parked at a predetermined parking position; an image information storage means 44 of storing the suitable parking images while relating them to own vehicle position information representing the current position of the own vehicle at the imaging; an image collating means 45 of collating a suitable parking image extracted as a reference image with real-time captured images acquired by the imaging means when it is decided that the vehicle is parked at the predetermined parking position based upon the own vehicle position information; and a stop position decision means 46 of deciding a travel-directional stop position of the vehicle, based on the collation result. <P>COPYRIGHT: (C)2009,JPO&INPIT

Description

  The present invention relates to a parking stop support system that performs support for accurately stopping a vehicle at an appropriate parking position during parking.

  There is a parking support system that supports a driver's parking operation with respect to a vehicle by displaying a captured image obtained by capturing a rear visual field of the vehicle during parking. As such a parking assistance device, the image is displayed based on a signal from a steering rudder angle sensor and a technology for assisting the driver's parking operation with respect to the vehicle by displaying an image of the situation in the backward direction with respect to the vehicle. A technique for assisting a driver's parking operation with respect to a vehicle by displaying a vehicle travel prediction curve superimposed on a camera image is known. At that time, a parking area line frame and a contour area of a car stop, which are generally photographed with a white line and photographed by a camera, are regarded as a parking area and are used for parking assistance.

  As one of such conventional techniques, when it is determined that the parking lot line imaged by the rear camera is incomplete when parking, it is stored in advance in an image database storing past captured images. Including the complete parking lot line by extracting the shot image that shows the same parking lot line from the past shot images that have been taken, and converting the viewpoint according to the current vehicle and superimposing it on the current rear camera shot image An apparatus for displaying a rear camera photographed image on a monitor is known (see Patent Document 1). In this device, even if the imaged parking lot line is incomplete due to circumstances such as the surrounding conditions of the parking lot, a clear parking lot line is displayed on the monitor. The parking area to be parked can be confirmed satisfactorily. However, this technology cannot be used in a parking lot that originally has no parking section line. In addition, when parking while looking at the parking lot line displayed on the monitor screen, even if it is relatively easy to put a vehicle in the parking lot line, it is not possible to stop at an accurate stop position with respect to the direction of travel. It is still difficult for drivers who are unfamiliar with the vehicle, and if there is no car stop, there is a possibility of causing the vehicle to collide with some obstacle or dropping the wheel into a groove.

JP 2006-311299 (paragraph number 0013, FIG. 2)

  Although there are parking lots that are frequently parked, there are parking lots that do not necessarily function effectively with conventional parking assistance systems because there are no parking lots or parking stops. Examples of such a parking lot include a parking space in a home site, a vertical approach type parking space provided between a building and a road, a mountain parking space, and the like. SUMMARY OF THE INVENTION The present invention has been made in view of the above problems, and its purpose is a parking lot that is frequently used in spite of the fact that there is no vehicle stop or the like, and parking for reliably stopping the vehicle at a set parking position. It is to provide an emergency stop support system.

  A parking stop support system according to the present invention for achieving the above object includes own vehicle position information acquisition means for acquiring own vehicle position information indicating the current position of the own vehicle, and when the vehicle is properly parked at a predetermined parking position. An imaging unit that acquires a captured image of the vehicle periphery as an appropriate parking image, an image information storage unit that stores the appropriate parking image in association with the host vehicle position information at the time of shooting, and the vehicle information based on the host vehicle position information. When it is determined that the vehicle is parked at a predetermined parking position, the appropriate parking image is extracted as a reference image from the image information storage unit, and the extracted reference image and the real-time captured image acquired by the imaging unit are used. Image collating means for collating, and stop position determining means for determining a stop position in the traveling direction of the vehicle based on a collation result by the image collating means.

  According to this configuration, there are no parking lots or parking stops, so even if it is a parking lot that does not function effectively with the conventional parking assistance system, the parking lot is only once with the help of other people. When parked at the correct stop position, the captured image around the vehicle when properly parked is registered as a proper parking image, stored in the image information storage means, and at the next parking at the same parking lot The corresponding proper parking image is extracted from the image information storage means as a reference image, and this reference image is collated with a real-time captured image acquired at the time of parking, and the coincidence of the collation result is the stop position of the vehicle, in particular It can be regarded as reaching the stop position in the traveling direction. Therefore, even in a parking lot without a parking lot or a car stop, the vehicle can be parked at an accurate stop position without the help of other people.

  Here, it is preferable that image processing means for performing image processing on the captured image acquired by the imaging means is provided, and an area selection process for deleting an area unnecessary for vehicle stop assistance in the captured image is performed. For example, when an appropriate parking image includes an image that exists only at the time of shooting, this becomes an obstacle at the time of comparison with a real-time shooting image that is performed later. For this reason, deleting a region that is not necessary for vehicle stop assistance, that is, matching between a proper parking image and a real-time captured image leads to an improvement in accuracy of the matching process.

  When it stops again in the parking lot which has already registered the appropriate parking image, this time may have stopped more correctly. In addition, due to seasonal fluctuations and the like, the captured image around the vehicle at the time of the current parking may be more suitable as an appropriate parking image. In order to answer such a problem, in a preferred embodiment of the present invention, the actual image acquired by the imaging unit when the appropriate parking at the predetermined parking position is completed through the stop position determination by the stop position determination unit. A configuration is adopted in which the time-captured image is stored as a new appropriate parking image in the image information storage means instead of the appropriate parking image used as the reference image.

  Recent automobiles and the like are equipped with a back camera, a front camera, or both as photographing means for preventing danger and driving support. In such a case, the captured image captured by each camera can be used as an appropriate parking image and a real-time captured image in the present invention. Therefore, when the imaging means includes a front camera whose subject is the front peripheral region of the vehicle, a rear camera whose subject is the rear peripheral region of the vehicle, or both, An appropriate parking image is stored and used as a later reference image and a real-time captured image.

  Furthermore, when both the back camera and the front camera are equipped, each camera can have a different role. For example, when the imaging means includes a front camera whose subject is the front peripheral region of the vehicle and a rear camera whose subject is the rear peripheral region of the vehicle, the captured image acquired by the rear camera during reverse parking The vehicle traveling direction stop position is determined from the comparison result with the corresponding reference image, and the vehicle traveling direction stop position is determined from the comparison result between the captured image acquired by the front camera and the corresponding reference image during forward parking. Can be configured to be determined. When parking in a mountain village parking lot where there is no parking stop, there are often cliffs and obstacles ahead of the parking direction, and whether it is forward parking or backward parking, it is positioned ahead of the parking direction The above embodiment is advantageous because the boundary line to be an important stop index.

  Parking stop support with sufficient time to prevent overshooting at the time of parking requires parking support based on the result of collating the reference image with the real-time captured image every moment. For this reason, it is convenient that the real-time captured images are sequentially acquired in time series by the imaging unit and sequentially verified with the reference image by the image matching unit.

  In addition, the image matching means includes a rough matching processing means for matching with the real-time photographed image to check whether the extracted reference image is appropriate, and a reference image deemed appropriate It is preferable to employ a configuration including precise matching processing means for matching pixel by pixel between the image and the real-time captured image. As a result, when the vehicle is away from the parking stop position, stop assistance is not necessary, so whether the reference image actually corresponds to the real-time photographed image by the coarse matching process with low calculation load and high-speed calculation. It is better to check the pre-matching process and the lateral displacement of the vehicle. When the vehicle approaches the stop position, the speed of the vehicle is slow. Therefore, it is preferable to provide reliable stop support through accurate precise matching even if the calculation speed is somewhat slow.

  Even if the parked position is substantially the same, the image characteristics (histogram characteristics, color balance, etc.) of the captured image vary depending on the shooting conditions, for example, the shooting time, the weather condition at the time of shooting, and the shooting season. Although fluctuations due to these shooting conditions can be covered to some extent by image processing, it is also convenient to register an appropriate parking image for each of these shooting conditions. For this reason, it is good to employ | adopt the structure by which the some suitable parking image from which imaging conditions differ in the same own vehicle position is stored in the said image information storage means as a suitable parking image candidate as a suitable form.

  In addition, the accuracy of the vehicle position information obtained by the vehicle position information acquisition unit is also somewhat accurate (for example, there is a possibility that an error of several meters may occur. For this reason, the image matching unit is configured to use the vehicle position information based on the vehicle position information. It is convenient to extract an appropriate parking image related to the position around the position from the image information storage means as an appropriate parking image candidate for the own vehicle position. The appropriate parking image registered at the peripheral position of the vehicle position information can also be used as an appropriate parking image candidate for this stop support.

  In the case of a parking lot without a car stop or the like, there are many cases where a lateral displacement of the vehicle is allowed to some extent. Therefore, in this parking stop support, it is also important to stop the vehicle at an appropriate stop position while allowing a certain amount of lateral displacement of the vehicle. For this reason, the amount of lateral displacement of the vehicle is determined from the collation result, and the image collating means allows the positional deviation below the predetermined amount in the lateral direction of the vehicle by performing collation in consideration of the positional displacement amount. It is also suitable to do.

  In the parking stop support system according to the present invention, it is important that stop position information generating means for generating stop position information for guiding the vehicle to the stop position based on the determination result by the stop position determining means is provided. It is. Based on the stop position information generated by the stop position information generating means, the sound of the content is played from the speaker, the message is displayed on the monitor, and the stop control of the vehicle itself can be performed from the stop position information. It becomes possible.

  In addition, the registration of the appropriate parking image in the present invention is based on the premise that parking in the parking lot will be performed again in the future. Therefore, a configuration in which the captured image acquired by the imaging unit when a predetermined registration operation is performed by a user (driver or his family) is stored in the image information storage unit as the appropriate parking image is convenient. is there.

  Even if the photographed image taken by the photographing means such as a camera is the same subject in the same place, the pixel density value constituting the picked-up image differs depending on various conditions such as cloudiness, backlight, and forward light. . However, the problem can be suppressed to some extent by normalizing in some form. As the simplest normalization, it is proposed to normalize by the average density value of the image.

  The technical features of the parking stop support system according to the present invention described above can also be applied to a parking stop support method and a parking stop support program. Therefore, the present invention also includes such a method and program. Can be targeted. For example, the parking stop support program for a vehicle according to the present invention comprising an imaging means for acquiring a captured image around the vehicle when the vehicle is properly parked at a predetermined parking position as the appropriate parking image. A function of storing in the image information storage means in association with own vehicle position information for acquiring own vehicle position information representing the current position of the own vehicle at the time of shooting, and determining that parking is performed at the predetermined parking position based on the own vehicle position information When the image is stored, the appropriate parking image is extracted from the image information storage means as a reference image, and the image reference function for comparing the extracted reference image and the real-time captured image acquired by the imaging means; The computer realizes a stop position determination function for determining a stop position in the traveling direction of the vehicle based on the result of verification by the image verification function. Naturally, such a parking stop support program can also obtain the effects described in the parking stop support system described above, and can also incorporate some additional techniques described as preferred examples thereof. It is.

  Hereinafter, an embodiment of a vehicle stop support system of the present invention incorporated in an automobile driving support system will be described in detail with reference to the drawings. FIG. 1 shows a schematic block diagram of the driving support system. This driving support system includes a back camera 1a as an imaging means for photographing a vehicle rear area, a navigation controller 2, a driving condition detection controller 3 that is a core element of the vehicle stop support system of the present invention, and a parking stop support controller. 4, a back guide monitor controller 5 and a notification means 6 are included. The notification means 6 includes a speaker 61 for outputting audio information and a monitor 62 for outputting image information.

  As shown in FIG. 2, the rear camera 1a is located near the upper center of the license plate mounted on the rear side of the vehicle in order to capture a region that is difficult to see from the driver's seat as a captured image when the vehicle is moved back to a predetermined location. In addition, it can be attached with the shooting axis direction slightly below the horizontal. Here, a camera using a CCD or a CMOS that captures one image (one frame) at a period of about 30 Hz can be applied. FIG. 2 shows a photographing area viewed from the lateral direction of the vehicle by the back camera 1a. The photographing area has a width of several meters from the rear of the vehicle and a width of about twice the vehicle body width. .

  The navigation controller 2 responds to signal inputs from a GPS (Global Positioning System) 22, a vehicle speed sensor that outputs vehicle speed pulses, a gyro sensor 23 that outputs direction information, an acceleration sensor 24 that outputs acceleration information, and the like. The vehicle position information generating means 20 for generating the vehicle position information and the map database 21 are provided. The navigation controller 2 monitors, as a guide image, an image obtained by superimposing the map information stored in the host vehicle location information map database 21 and a sign indicating the host vehicle based on the information acquired by the host vehicle location information generation unit 20. To display.

  The driving condition detection controller 3 is connected to sensors such as a vehicle speed sensor 31, a steering angle sensor 32, a shift position detection sensor 33, a brake sensor 34, and an accelerator sensor 35. The vehicle speed sensor 31 is composed of an active wheel speed sensor attached to a wheel of the vehicle, detects the rotational speed of the wheel, and outputs it as a speed signal. The steering angle sensor 32 is attached to the inside of the steering device, detects the steering angle when the steering wheel is steered, and outputs it as a steering angle signal. The shift position detection sensor 33 is built in the shift lever, and the shift position is “P (parking)”, “N (neutral)”, “R (reverse)”, “D (drive)”, “2 (second speed)”. , “L (low)” is detected. The brake sensor 34 detects the driver's brake operation amount, and the accelerator sensor 35 detects the driver's accelerator operation amount. The driving situation detection controller 3 can also determine a driving situation of the vehicle, for example, a reverse running situation other than the normal running situation, a parking running situation, and the like based on signal inputs from these sensor groups.

  Based on the vehicle position information from the vehicle position information generation unit 20 and the determination of the parking travel status of the driving condition detection controller 3, the parking stop support controller 4 moves the driver to the predetermined parking position of the predetermined parking lot. Assist in stopping properly. The parking stop support controller 4 is a core element of the parking stop support system according to the present invention, and a detailed configuration thereof will be described later in detail.

  The back guide monitor controller 5 supports a safe reverse travel for the driver by displaying, on the monitor 62, a captured image of a peripheral area behind the vehicle in the traveling direction acquired by the back camera 1a during reverse travel. To do.

 As shown in FIG. 3, the parking stop support controller 4 inputs own vehicle position information from the own vehicle position information generating unit 20 of the navigation controller 2, inputs a photographed image from the back camera 1a, and includes a touch panel and various buttons. An operation command is input from the operation input unit 7 comprising: The notification unit 6 also functions as an output unit of the parking stop support controller 4 and provides guidance to the driver until the vehicle reaches a predetermined stop position or at least a speaker 61 or a monitor 62 that the vehicle has reached the predetermined stop position. Notify through. It is also possible to transmit to the vehicle control system that the vehicle has reached a predetermined stop position and use it for vehicle control (brake operation, etc.).

  This parking stop support controller 4 refers to an appropriate parking image storage mode in which a captured image around the vehicle when the vehicle is properly parked at a predetermined parking position is stored as an appropriate parking image, and the appropriate parking image is referred to at the next parking. It has an appropriate parking support mode for guiding the appropriate parking of the vehicle as an image. The determination of these modes is performed by the parking stop management means 40 in response to the input vehicle position information, the signal indicating that the driving condition detection controller 3 has determined the parking traveling condition, and the operation command from the operation input unit 7. Do. The parking stop management means 40 also performs control management of various components of the parking stop support controller 4 described below.

  In the proper parking image storage mode, a captured image of the rear peripheral area by the back camera 1a when the vehicle is properly parked at a predetermined parking position is input to the image processing means 41. The image processing means 41 performs an outline emphasis process or the like on the input photographed image so as to facilitate the subsequent collation process with other photographed images. The appropriate parking image processed by the image processing means 41 is transferred to the appropriate parking image information creation unit 42. Since the appropriate parking image information creation unit 42 receives the current vehicle position information through the own vehicle position information acquisition means 43, the appropriate parking image information is created by associating the own vehicle position information with the appropriate parking image. The created appropriate parking image information is stored in the image information storage means 44 so that it can be searched and extracted by the vehicle position information. The stored appropriate parking image information is extracted as a reference image for appropriate parking at the same parking position in the future.

  In the appropriate parking support mode that is performed when it is determined that the appropriate parking image has been parked at the parking position stored in the image information storage unit 44 and the appropriate parking image at the stop position has been previously stored, the image verification unit 45 Do the central work. The image matching means 45 searches the image information storage means 44 using the own vehicle position information of the vehicle that is about to be parked as a search condition, and develops the appropriate parking image from the matched appropriate parking image information as a reference image in the working memory. When the proper parking support mode is entered, the captured image acquired by the back camera 1a is transferred to the image processing means 41 as a real-time captured image as needed. The real-time photographed image is sent to the image collating unit 45 after the image processing unit 41 performs the same contour enhancement processing as that of the previous appropriate parking image.

  The image matching unit 45 performs matching between the real-time photographed image and the reference image by using an image matching method such as pattern matching. The image matching unit 45 and the coarse matching unit 45a that performs rough matching and the fine matching unit that performs pixel-based precise matching. And a precision collating means 45b.

  The coarse matching unit 45a checks whether the reference image developed in the working memory is appropriate as a matching source in the current proper parking, so that a region that matches a partial region of the real-time captured image becomes a reference image. Evaluate if it exists. If the vehicle is parked, the real-time captured image will be sent sequentially in time series because the shooting point of the real-time captured image approaches the shooting point of the reference image as the vehicle moves. This rough collation is performed for. However, if there is no region in the reference image that matches a partial region of the real-time captured image despite a certain amount of vehicle travel, the next reference image candidate is developed in the working memory and this coarse matching is repeated. It is also possible. When a plurality of appropriate parking images with different shooting conditions at the same vehicle position are stored in the image information storage means 44, the next appropriate parking image is designated as a reference image candidate. Alternatively, when the appropriate parking image associated with the peripheral position of the current vehicle position is stored in the image information storage means 44, the appropriate parking image at this peripheral position is designated as a reference image candidate.

  If there is an area in the reference image that matches a partial area of the real-time photographed image, the collation process is taken over by the fine collation means 45b. The precise matching process is performed until the real-time captured image and the reference image that are sequentially sent coincide with each other in the vehicle traveling direction. At that time, the lateral position deviation amount in the vehicle traveling direction between the photographing point of the appropriate parking image as the reference image and the photographing point of the real time photographed image is detected, and the real time photographed image and the reference image corresponding to the lateral position deviation amount are detected. Shift one of them. Thereby, it is possible to allow a positional shift of about the width of the vehicle in the verification process. That is, in actual parking, an accurate stop line in the traveling direction can be protected even if the parking posture is shifted by about 1 m in the lateral direction, for example.

  A pixel shift amount in the vehicle traveling direction, which is a result of matching between the real-time captured image and the reference image by the precise matching unit 45b, is sent to the stop position determining unit 46. The stop position determination means 46 calculates the distance to the stop line prescribed | regulated by the appropriate parking image used as a reference image from the pixel shift amount of the vehicle advancing direction sent. The vehicle may be stopped at a position where the distance to the stop line becomes zero. The distance to the stop line determined by the stop position determination means 46 is sent to the stop position information generation means 47. The stop line position information generation unit 47 generates the received distance to the stop line as display information or audio information, and outputs it to the notification unit 6. As the display information, a mode in which the distance to the stop line is simply represented by characters may be adopted, but a symbol such as a bar as a stop line is superposed on the real-time captured images sequentially sent from the camera 1. It is convenient to adopt an imposing mode.

An appropriate parking image registration control routine in the parking stop support controller 4 configured as described above will be described with reference to the flowchart of FIG.
In the first parking lot, parking is performed while looking at an image behind the vehicle projected on the monitor 62 by the back guide monitor controller 5 or through guidance by an assistant. Completion of the parking (stopping of the reverse vehicle) is determined by the driving situation controller 3. When the driving status controller 3 notifies the parking stop support controller 4 of the completion of parking (# 10 Yes branch), the parking stop management means 40 notifies the driver whether or not to register an appropriate parking image related to this parking. If the driver does not want the registration (# 11 No branch), this routine ends.

  If the driver wishes to register (# 11 Yes branch), the image captured by the back camera 1a is acquired as an appropriate parking image (# 12). The acquired proper parking image is subjected to predetermined image processing by the image processing means 41 (# 13). This image processing is for facilitating later image collation, and level adjustment, contour enhancement, posterization, etc. are employed. At the time of image collation, it is preferable to convert to a binarized image in order to avoid a calculation burden, but a color image or a grayscale image may be used when registering an appropriate parking image. However, since the pixel density value of the captured image is different depending on the shooting location, the shooting time, the weather, etc., it is convenient to normalize using the average density value when entering the image processing means 41. It is.

  When image processing as pre-processing for proper parking image registration is completed, the vehicle position information acquisition unit 43 acquires the vehicle position information from the vehicle position information generation unit 20 of the navigation controller 2 (# 14). The acquired own vehicle position information and the appropriate parking image for which the image processing has been completed are sent to the appropriate parking image information creating means 42, where the appropriate parking image and the own vehicle position information are associated with each other and created as appropriate parking image information. (# 15). The appropriate parking image information creation means 42 accesses the image information storage means 44 that stores appropriate parking image information at various parking positions, and is substantially the same as the created appropriate parking image information. It is checked whether or not appropriate parking image information having (parking position) is stored (# 16). Here, substantially the same vehicle position information means vehicle position information included in a predetermined range set in consideration of a measurement error of the vehicle position information, an allowable error of the parking position, and the like. This predetermined range may be fixedly set, but may be variably set depending on the error situation of the vehicle position information, the characteristics of the parking lot (the size of the parking space, the surrounding situation), and the like.

  When proper parking image information having the same vehicle position information (parking position) as the created appropriate parking image information is not stored (# 16 No branch), the created appropriate parking image information is newly registered as appropriate parking image information. Is stored in the image information storage means 44 (# 17). When proper parking image information having the same own vehicle position information (parking position) as the created appropriate parking image information is stored (# 16 Yes branch), the appropriateness included in the already registered appropriate parking image information It is checked whether the image characteristics of the parking image and the appropriate parking image of the appropriate parking image information that has been created are substantially the same (# 18). This is because even if the photographed image is normalized as described above, there are different image characteristics even in the photographed image at the same photographing point due to the difference in the photographing conditions, especially the subject that varies with time and season. Because it is shown. When the image features of both are not substantially the same (# 18 No branch), the appropriate parking image acquired this time can be a candidate of an appropriate parking image at the same shooting point, and the appropriate parking image information is an additional registration image. It is stored in the information storage means 44 (# 19). When both image features are substantially the same (# 18 Yes branch), the newly created appropriate parking image information is assumed to represent the latest surrounding situation of the stop position, and the old appropriate parking image information is updated. In order to do this, the latest appropriate parking image information is stored in the image information storage means 44 as rewrite registration (# 20).

Next, the driving assistance record control routine will be described with reference to the flowchart of FIG.
Although the driving state detection process of the vehicle by the driving detection controller 3 is performed, if the pre-parking state where the vehicle is about to park is not detected (# 51 No branch), a driving support routine for normal driving is performed. When the driving detection controller 3 detects the pre-parking situation (# 51 Yes branch), further, based on information such as the traveling speed, the steering wheel turning operation, and the forward / backward shift switching, the actual parking situation is implemented. Whether it has been reached is checked (# 52). If the parking execution status is not detected within a predetermined time after the detection of the pre-parking status (# 53 Yes branch), it is determined that the parking status is not reached, and the process returns to step # 51. When the driving detection controller 3 detects the parking implementation status (# 52 Yes branch), the possibility of parking support by the parking stop support controller 4 is determined.

  First, the vehicle position information acquisition means 43 acquires the vehicle position information from the vehicle position information generation means 20 of the navigation controller 2 (# 61). Using the acquired own vehicle position information as a search condition, the image information storage means 44 is searched for appropriate parking image information that matches (# 62). If there is no appropriate parking image information suitable for the image information storage means 44 (# 63 No branch), it is determined that parking stop support using an appropriate parking image is impossible, and a normal parking support routine, for example, a back guide monitor controller 5 is used to provide parking assistance for displaying a captured image of the rear of the vehicle on the monitor 6.

  When appropriate parking image information suitable for the image information storage means 44 exists (# 63 Yes branch), the parking stop support routine using the registered appropriate parking image by the parking stop support controller 4 is started. First, the appropriate appropriate parking image information is extracted from the image information storage means 44, and the appropriate parking image is developed in the working memory (# 70). At that time, when there are a plurality of suitable appropriate parking image information, all of them are extracted, and each appropriate parking image is developed as a proper parking image candidate in the working memory. Considering the problem of the accuracy of the own vehicle position information, it is preferable that the appropriate parking image information related to the own vehicle position information within a predetermined range is also the appropriate parking image candidate. Furthermore, the image captured by the back camera 1a is acquired as a real-time parking image (# 71). The acquired real-time parking image is subjected to the same processing as the image processing performed by the image processing means 41 when the proper parking image is registered (# 72). FIG. 6 shows a proper parking image binarized after image processing such as normalization, level adjustment, contour enhancement, posterization, etc. (FIG. 6A) and real-time captured image (FIG. 6B). Are illustrated respectively.

  This proper parking image is set as a reference image for image matching (image matching) (# 73), and rough matching is first performed (# 74). In the rough matching, since it is checked whether or not the set proper parking image can be used as a reference image, as illustrated in FIG. 7, the simplified appropriate parking image (FIG. 7A) and the real-time captured image are used. Using (FIG. 7 (b)) is also advantageous from the point of calculation load. In addition, since the shooting point of the real-time shooting image acquired at this time is on the near side of the shooting point of the appropriate parking image at the vehicle proper stop position, it is collated with a similar shape or both image features The comparison may be performed after applying appropriate image enlargement to the real-time photographed image. If no match between the proper parking image and the real-time photographed image is recognized in this rough collation (# 75 No branch), it is checked whether or not the next candidate for the appropriate parking image is developed in the working memory (# 76). If there is no next candidate for the appropriate parking image (# 76 No branch), it is determined that the parking stop support using the appropriate parking image is impossible, and the routine proceeds to a normal parking support routine. When there is a next candidate for a proper parking image (# 76 Yes branch), a fine matching process for matching pixel by pixel between the reference image and the sequentially acquired real-time captured image is started. Before performing the fine matching process, the lateral displacement amount of the vehicle is determined from the coarse matching result, and the lateral displacement amount is set and used in the precise matching process (# 77). In other words, by shifting the relative position of the reference image and the real-time captured image in the left-right direction of the vehicle in the precise matching process by the amount of the position shift, the position shift in the left-right direction of the parked vehicle by the parking stop support is correspondingly increased. Only, that is, a positional deviation of a predetermined amount or less is allowed.

  In the precise matching process, first, the latest real-time photographed image is acquired from the back camera 1a (# 80). The acquired real-time captured image (for example, FIG. 6 (b)) and the appropriate parking image (for example, FIG. 6 (a)) set as the reference image are collated on a pixel basis (# 81). The collation in units of pixels includes collation performed by blocking several hundred pixels into one pixel. As a result of this precise matching, it is determined how many pixels the real-time captured image is shifted in the vehicle traveling direction with respect to the appropriate parking image. Therefore, the distance from the shifted pixel to the stop position is the stop position determination means 46. Is calculated as stop position determination (# 82). The distance to the stop position determined by the stop position determination unit 46 is converted into notification information by the stop position information generation unit 47 (# 83), and is notified through the speaker 61 or the monitor 62 of the notification unit 6 as necessary. (# 84). As long as the distance to the stop position does not become substantially zero (# 85 No branch), the process returns to step # 80, and the precise collation using the latest real-time photographed image acquired from the back camera 1a is repeated. If the distance to the stop position becomes substantially zero, the vehicle has reached an appropriate stop position, and a notification to that effect is given. When the vehicle reaches an appropriate stop position (# 85 Yes branch), a configuration may be adopted in which the vehicle is automatically stopped. This routine ends when the vehicle stops.

  Although not shown in this flowchart, if it is desired to register an appropriate parking image at this stop position after the vehicle has stopped, steps # 12 to # 16 in the flowchart shown in FIG. After that, it is possible to perform additional registration or rewrite registration of the appropriate parking image.

[Another embodiment]
(1)
In the embodiment described above, the reverse parking using only the image captured by the back camera 1a has been described. However, the forward parking can be similarly performed using only the image captured by the back camera 1a. In this case, when registering an appropriate parking image, a code or the like for distinguishing forward parking or reverse parking is associated with the vehicle position information. When parking stop support is performed, it is determined whether the parking to be carried out is forward parking or reverse parking, and in the case of forward parking, appropriate parking information during forward parking is stored as image information storage means. It is good to extract from 44 and use for a collation process.

(2)
As another embodiment, as shown in FIG. 8, a back camera 1a and a front camera 1b are mounted on a vehicle, and an image captured by the back camera 1a is used during reverse parking, and an image captured by the front camera 1b is used during forward parking. It is also possible. In parking lots in private homes and in mountain villages, there are many targets that are relatively easy to detect their contours, such as obstacles and cliffs, in the direction of parking. For this reason, a photographed image with the subject in the parking direction is suitable as the appropriate parking image. The difference between the flow of driving parking image registration control shown in FIG. 9 in this embodiment and the flow of driving parking image registration control shown in FIG. 4 in the first embodiment is that a captured image is acquired as a proper parking image. (Step # 12 in FIG. 4). In the first embodiment, since only the photographed image by the back camera 1a is used, the photographed image acquired as the proper parking image is the one captured by the back camera 1a. However, in this embodiment, before the step of acquiring the captured image as the appropriate parking image, that is, when the driver desires to register the appropriate parking image (# 11 Yes branch), the parking direction is determined (# 12a). The determination of the parking direction is performed based on a code indicating a reverse parking situation or a forward parking situation, which is returned by making an inquiry to the driving situation detection controller 3. When it is determined that the parking direction is reverse parking, a captured image by the back camera 1a is acquired as an appropriate parking image (# 12b). If it is determined that the parking direction is forward parking, an image captured by the back camera 1b is acquired as an appropriate parking image (# 12c). In the appropriate parking image information creation process in step # 15, the appropriate parking image information is created by associating the appropriate parking image with the vehicle position information and an identifier for distinguishing forward parking or backward parking.

  Further, the difference between the flow of the driving support control shown in FIG. 10 in this embodiment and the flow of the driving support control shown in FIG. 5 in the first embodiment is a step of searching for an appropriate parking image ( Step # 62) in FIG. In the first embodiment, the proper parking image is only associated with the vehicle position information and stored in the image information storage means 44, and therefore the search condition is only the vehicle position information. However, in this embodiment, the appropriate parking image is associated with the own vehicle position information and the parking direction, and therefore the own vehicle position information and the parking direction are required as the search conditions. Therefore, the parking direction is determined before searching for an appropriate parking image (# 62a). The determination of the parking direction is performed based on a code indicating a reverse parking situation or a forward parking situation, which is returned by making an inquiry to the driving situation detection controller 3. When it is determined that the parking direction is reverse parking, an appropriate parking image is searched using the identifier indicating reverse parking and the vehicle position information acquired in step # 61 as search conditions (# 12b). When it is determined that the parking direction is forward parking, an appropriate parking image is searched using the identifier indicating forward parking and the own vehicle position information acquired in step # 61 as search conditions (# 12c). The other processes are substantially the same as those in the first embodiment.

  That is, in this embodiment, the imaging means includes the front camera 1b whose subject is the front peripheral region of the vehicle and the rear camera 1a whose subject is the rear peripheral region of the vehicle, and is acquired by the rear camera during reverse parking. The stop position of the vehicle in the advancing direction is determined from the collation result between the captured image and the corresponding reference image, and the vehicle progresses from the collation result between the captured image acquired by the front camera and the corresponding reference image during forward parking. A direction stop position is determined.

(3)
When the back camera 1a and the front camera 1b are mounted on the vehicle, and an image captured by the back camera 1a and an image captured by the front camera 1b can be used, an embodiment using both the captured images as appropriate parking images is also possible. It is. At that time, if the matching process is performed for each of the appropriate parking images and the stop position is determined based on the matching result with the higher matching accuracy, the reliability can be improved.

(4)
Parking that requires assistance for the stop position during parking is generally reverse parking. Considering this, when both the back camera 1a and the front camera 1b can be used, the image captured by the back camera 1a is used as the main proper parking image, and the image captured by the front camera 1b is used as the sub proper parking. Embodiments used as images are also possible. The main appropriate parking image is handled in the same manner as in the first embodiment, and the sub appropriate parking image may be additionally used in the rough matching. For example, the sub proper parking image related to the main proper parking image and the real-time captured image sent from the front camera 1b are roughly checked to determine whether a large error has occurred in the lateral direction of the vehicle. Can be determined. At that time, the photographing field of view of the front camera 1b is set larger than that of the back camera 1a so that a photographed image suitable for rough position judgment can be obtained even though it is not suitable for precise stop position judgment. Convenient.

(5)
In the description of the above-described embodiment, the image processing unit 44 performs image processing such as level adjustment, contour enhancement, and posterization so that the received captured image is suitable for image collation performed later. However, in order to efficiently perform the image matching process, it is also one of preferred embodiments that the image processing unit 44 performs a region selection process for deleting a region unnecessary for the vehicle stop support in the captured image. For example, other cars such as cars and dogs that exist only at the time of shooting in the shot image are to be deleted. When the deletion target is moving, a pixel area in which a difference value is generated between two captured images acquired at a predetermined time interval can be regarded as a deletion area or a verification unnecessary area. In addition, in the case of a deletion target having a specific shape such as a car, it is recognized that the object is a car-like object from the contour shape and density value information, and the pixel area is regarded as a deletion area or a verification unnecessary area. Is also possible.

1 is a block diagram of an automobile driving support system incorporating a vehicle stop support system according to the present invention. Schematic diagram showing the shooting area of the back camera attached to the car Functional block diagram of the vehicle stop support system Flow chart of appropriate parking image registration control Driving support control flowchart It is a figure which shows an example of the picked-up image by which the image process was carried out, (a) is a figure which shows an example of a proper parking image, (b) is a figure which shows an example of a real-time picked-up image. It is a figure which shows an example of the picked-up image simplified in the shape of the substantially line figure, (a) is a figure which shows an example of a proper parking image, (b) is a figure which shows an example of a real-time picked-up image. Schematic diagram showing the shooting area of the back camera and front camera attached to the car Flowchart of proper parking image registration control in another embodiment Flowchart of driving support control in another embodiment

Explanation of symbols

1a: Back camera (imaging means)
1b: Front camera (imaging means)
3: Driving status detection controller 4: Parking stop support controller 6: Notification unit 7: Operation input unit 20: Own vehicle position information generation unit 40: Parking stop management unit 41: Image processing unit 42: Appropriate parking image information generation unit 43: Own vehicle position information acquisition means 44: Image information storage means 45: Image collation means 45a: Coarse collation means 45b: Fine collation means 46: Stop position determination means 47: Stop position information generation means

Claims (14)

Own vehicle position information acquisition means for acquiring own vehicle position information representing the current position of the own vehicle;
Imaging means for acquiring a captured image around the vehicle when properly parked at a predetermined parking position as an appropriate parking image;
Image information storage means for storing the appropriate parking image in association with the vehicle position information at the time of shooting;
When it is determined that the vehicle is parked at the predetermined parking position based on the vehicle position information, the appropriate parking image is extracted as a reference image from the image information storage unit, and the extracted reference image and the imaging unit are used. Image collating means for collating the acquired real-time captured image;
Stop position determination means for determining a stop position in the traveling direction of the vehicle based on the result of verification by the image verification means;
A parking stop support system. The parking process according to claim 1, further comprising image processing means for performing image processing on the captured image acquired by the imaging means, and performing area selection processing for deleting an area unnecessary for vehicle stop assistance from the captured image. Stop support system. Appropriate parking used as the reference image, using the real-time captured image acquired by the imaging means as a new appropriate parking image when appropriate parking at a predetermined parking position is completed through stop position determination by the stop position determining means The parking stop support system according to claim 1 or 2, wherein the parking information is stored in an image information storage means instead of an image. The imaging means includes a front camera whose subject is the front peripheral region of the vehicle, a rear camera whose subject is the rear peripheral region of the vehicle, or both, and the appropriate parking image is stored for each camera provided. The parking stop support system according to any one of claims 1 to 3, which is used as a later reference image and a real-time captured image. The imaging means includes a front camera whose subject is the front peripheral region of the vehicle and a rear camera whose subject is the rear peripheral region of the vehicle, and a captured image acquired by the rear camera during reverse parking and a reference corresponding thereto The vehicle traveling direction stop position is determined from the collation result with the image, and the vehicle traveling direction stop position is determined from the collation result between the captured image acquired by the front camera and the corresponding reference image during forward parking. The parking stop support system according to any one of Items 1 to 3. The parking stop support system according to any one of claims 1 to 5, wherein the real-time captured images are sequentially acquired in time series by the imaging unit and sequentially compared with the reference image by the image collating unit. The image matching means includes a coarse matching processing means for matching with the real-time photographed image to check whether the extracted reference image is appropriate, a reference image deemed appropriate, and the The parking stop support system according to any one of claims 1 to 6, further comprising a fine matching processing unit that performs pixel-by-pixel matching with a real-time captured image. The parking stop support system according to any one of claims 1 to 7, wherein a plurality of appropriate parking images having different shooting conditions at the same vehicle position are stored in the image information storage unit as appropriate parking image candidates. The said image collation means extracts the appropriate parking image linked | related with the peripheral position of the own vehicle position based on the said own vehicle position information from the said image information storage means as an appropriate parking image candidate of the said own vehicle position. The parking stop support system according to any one of 1 to 8. The lateral displacement amount of the vehicle in the left-right direction is determined from the collation result, and the image collating unit allows a positional displacement equal to or less than a predetermined amount in the lateral direction of the vehicle by performing collation in consideration of the positional displacement amount. The parking stop support system according to any one of 1 to 8. The parking according to any one of claims 1 to 9, further comprising stop position information generating means for generating stop position information for guiding the vehicle to a stop position based on a determination result by the stop position determining means. Time stop support system. The parking according to any one of claims 1 to 11, wherein a captured image acquired by the imaging unit when a predetermined registration operation is performed by a user is stored in the image information storage unit as the appropriate parking image. Time stop support system. The parking stop assistance system according to any one of claims 1 to 12, wherein the reference image and the real-time captured image that are collated by the image collating unit are normalized by an average density value. In a parking stop support program for a vehicle provided with an imaging means for acquiring a captured image around the vehicle when appropriately parked at a predetermined parking position as an appropriate parking image,
A function of storing the appropriate parking image in image information storage means in association with own vehicle position information for acquiring own vehicle position information representing the current position of the own vehicle at the time of shooting;
When it is determined that the vehicle is parked at the predetermined parking position based on the vehicle position information, the appropriate parking image is extracted as a reference image from the image information storage unit, and the extracted reference image and the imaging unit are used. An image matching function that matches the acquired real-time image,
A stop position determining function for determining a stop position in the traveling direction of the vehicle based on a collation result by the image collating function;
A parking stop support program to make a computer realize.

Images