JP2013015882A - Image recognition device for vehicle - Google Patents

Abstract

PROBLEM TO BE SOLVED: To provide an image recognition device for a vehicle which determines own vehicle periphery state using a single camera without moving own vehicle when starting the own vehicle.SOLUTION: The image recognition device includes: the camera for photographing the own vehicle periphery; stop time storage control means for causing storage means to store information showing a stop time state in the own vehicle periphery during own vehicle stop time (more specifically, picked-up images by the camera during the own vehicle stop time and the presence or absence and a position of an obstacle existing in a direction photographed by the camera during the own vehicle stop time); state determination means for determining the own vehicle periphery state on the basis of the picked-up images by the camera and the stop time state stored as information in the storage means when starting the own vehicle.

Description

  The present invention relates to a vehicular image recognition device, and more particularly to a vehicular image recognition device that is suitable for determining a state around a host vehicle by photographing the periphery of the host vehicle using a camera.

  2. Description of the Related Art Conventionally, a vehicular image recognition device that captures an image of the periphery of the host vehicle and detects a distance to an obstacle around the host vehicle is known (see, for example, Patent Document 1). This image recognition device uses a single camera mounted on the host vehicle to acquire two images in time series at two different points around the host vehicle and calculate vehicle movement data between the two points. The distance to the obstacle is calculated based on the principle of triangulation for the obstacle displayed in the two images. Therefore, according to the above-described image recognition apparatus, an obstacle existing around the host vehicle can be detected using a single camera.

JP 2001-187553 A

  However, in the apparatus described in Patent Document 1, since it is necessary for the host vehicle and the obstacle to move relative to each other in order to detect the obstacle using a single camera, the host vehicle starts from a stopped state. When the vehicle does not move the vehicle at a certain distance from the stop position, images at different points cannot be acquired, and as a result, an obstacle cannot be detected. Therefore, in the above-described method, it is necessary to move the host vehicle from the stop position in order to determine the state around the host vehicle using a single camera when the host vehicle starts.

  The present invention has been made in view of the above points, and is for a vehicle that can determine the state of the vicinity of the own vehicle using a single camera without accompanying the movement of the own vehicle when the own vehicle starts. An object is to provide an image recognition apparatus.

  The purpose of the above is to provide a camera for photographing the periphery of the host vehicle, a storage control unit for storing information indicating the stop state of the periphery of the host vehicle when the host vehicle is stopped, a storage control unit for stopping the vehicle, This is achieved by a vehicle image recognition apparatus comprising: a state determination unit that determines a state around the host vehicle based on the captured image and the stop state stored in the storage unit.

  Note that “when the host vehicle is stopped” may have a certain time span from immediately before the speed of the host vehicle becomes zero to immediately after it becomes zero. Further, “when the host vehicle starts” may include a time when the host vehicle is stopped immediately before starting to travel, but does not include a time after the host vehicle starts to travel from the stopped state. .

  According to the present invention, it is possible to determine the state around the host vehicle using a single camera without accompanying movement of the host vehicle when the host vehicle starts.

1 is a configuration diagram of an image recognition apparatus for a vehicle that is an embodiment of the present invention. It is a flowchart of an example of the control routine performed when the own vehicle stops in the vehicle image recognition apparatus of the present embodiment. It is a flowchart of an example of the control routine performed at the time of the own vehicle start in the image recognition apparatus for vehicles of a present Example. It is a figure for demonstrating the method of determining that there is no obstruction at the time of vehicle start, when there is no image change by the time of vehicle stop and the time of vehicle start in the situation where there was no obstruction at the time of vehicle stop. It is a figure for demonstrating the method of determining with an obstacle at the time of vehicle start, when there is no image change by the time of a vehicle stop and the time of a vehicle start in the situation where there was an obstacle at the time of a vehicle stop. It is a figure for demonstrating the method of determining with an obstacle at the time of vehicle start, when there exists an image change at the time of a vehicle stop and the time of a vehicle start in the condition where there was no obstacle at the time of a vehicle stop.

  Hereinafter, specific embodiments of a vehicle image recognition apparatus according to the present invention will be described with reference to the drawings.

  FIG. 1 shows a configuration diagram of a vehicle image recognition apparatus 10 according to an embodiment of the present invention. The vehicle image recognition apparatus 10 according to the present embodiment is mounted on a vehicle such as an automobile, and exists in the vicinity of the own vehicle (specifically, on the rear side of the vehicle) based on a captured image obtained with only a single camera. It is an image recognition device for detecting an obstacle as a three-dimensional object (for example, a person, a wall, a power pole, another vehicle, etc.). The vehicular image recognition device 10 may be applied to, for example, a bag guide monitor system.

  As shown in FIG. 1, the vehicular image recognition device 10 includes an imaging device 12. The imaging device 12 is a single digital camera configured by a solid-state imaging device such as a charge coupled device (CCD). Hereinafter, the imaging device 12 is referred to as a monocular camera 12. The monocular camera 12 is attached to the rear part of the vehicle body, and can capture a predetermined range extending from the rear part of the vehicle body to the rear of the host vehicle.

  The vehicular image recognition apparatus 10 also includes an electronic control unit (hereinafter referred to as an image processing ECU) 14 mainly composed of a microcomputer. The monocular camera 12 is electrically connected to the image processing ECU 14. An image captured by the monocular camera 12 is supplied to the image processing ECU 14. The supply of the captured image from the monocular camera 12 to the image processing ECU 14 is performed every predetermined time.

  Further, a steering angle sensor 16, a wheel speed sensor 18, and a shift position sensor 20 are electrically connected to the image processing ECU 14. The steering angle sensor 16 outputs a signal corresponding to the steering angle of the host vehicle to the image processing ECU 14. The wheel speed sensor 18 outputs a signal corresponding to the wheel rotation speed of the host vehicle to the image processing ECU 14. The shift position sensor 20 outputs a signal corresponding to the shift shift position of the host vehicle to the image processing ECU 14.

  The image processing ECU 14 detects the steering angle of the host vehicle based on the signal from the steering angle sensor 16. Further, the number of wheel rotations of the host vehicle is detected based on a signal from the wheel speed sensor 18, and the amount of movement of the host vehicle during a predetermined time is detected. Further, the shift shift position of the host vehicle is detected based on the signal from the shift position sensor 20. The image processing ECU 14 exists behind the host vehicle by performing image processing based on the captured image data supplied from the monocular camera 12 and the data supplied from the sensors 16 to 20 while the vehicle is traveling. The presence or absence of an obstacle as a three-dimensional object is determined, and when it is determined that the obstacle is present, the position or distance of the obstacle with respect to the host vehicle is further detected.

  A display monitor 22 and an alarm sound device 24 are also electrically connected to the image processing ECU 14. The display monitor 22 is disposed on the center panel or the like so as to be visible to the driver who drives the host vehicle. For example, the display monitor 22 may also be used as a monitor used in a navigation system. Further, the alarm sound device 24 is a vehicle interior buzzer, a speaker, or the like disposed so that a driver driving the host vehicle can hear.

  The image processing ECU 14 processes a captured image of the monocular camera 12 and outputs the processed image to the display monitor 22. The display monitor 22 displays an image supplied from the image processing ECU 14. The image displayed on the display monitor 22 may be an image captured by the monocular camera 12 converted into a bird's eye view image, and the position and distance of the obstacle detected as described above. Data to which data is added may be used.

  Further, when the image processing ECU 14 determines that there is an obstacle existing behind the host vehicle as described above, when the host vehicle approaches a predetermined distance or less with respect to the obstacle, the image processing ECU 14 warns the driver of the host vehicle. The audible alarm device 24 is activated to provide. The alarm sound device 24 generates an alarm sound for the driver of the host vehicle in accordance with an instruction from the image processing ECU 14. The alarm sound may be a buzzer sound or a voice instruction.

  The image processing ECU 14 includes image memories 30 and 32, a CPU 34, a RAM 36, and a ROM 38. An image captured by the monocular camera 12 is converted from an analog image to a digital image, and then supplied to the image memory 30. The image memory 30 stores digital image data from the monocular camera 12 and supplies it to the CPU 34. The ROM 38 stores a program for performing image conversion, a program for calculating a moving amount and a moving position of the vehicle, and the like, and appropriately develops the program in the RAM 36. The RAM 36 is a work area for executing the above-described program.

  The CPU 34 is based on the digital image data from the monocular camera 12, the steering angle data from the steering angle sensor 16, the wheel rotational speed data from the wheel speed sensor 18, and the shift shift position data from the shift position sensor 20. Then, image processing is performed according to the program on the RAM 36. Specifically, this image processing determines whether or not there is an obstacle as a three-dimensional object existing behind the host vehicle. When it is determined that there is an obstacle, the data for identifying the obstacle and the obstacle The position or the distance from the host vehicle to the obstacle is calculated, and each data is added to the digital image from the image memory 30, and the image is converted into a bird's eye view image.

  Bird's eye view image data obtained by image processing by the CPU 34 is supplied to the image memory 32. The image memory 32 stores the bird's eye view image data from the CPU 34 and outputs it to the display monitor 22. The display monitor 22 displays a bird's eye view image from the image processing ECU 14.

  Further, when the CPU 34 determines that there is an obstacle existing behind the host vehicle, the alarm sound device is used to give a warning to the driver of the host vehicle when the host vehicle approaches the obstacle below a predetermined distance. 24 is activated. The alarm sound device 24 generates an alarm sound for the driver of the host vehicle in accordance with an instruction from the CPU 34.

  Next, with reference to FIG. 2 thru | or FIG. 6, operation | movement of the image recognition apparatus 10 for vehicles of a present Example is demonstrated. FIG. 2 shows a flowchart of an example of a control routine executed by the image processing ECU 14 when the host vehicle is stopped in the vehicle image recognition apparatus 10 of the present embodiment. FIG. 3 shows a flowchart of an example of a control routine executed by the image processing ECU 14 when the host vehicle starts in the vehicle image recognition apparatus 10 of the present embodiment. FIG. 4 is a diagram for explaining a method for determining that there is no obstacle when the vehicle starts when there is no image change between when the vehicle stops and when the vehicle starts in a situation where there is no obstacle when the vehicle stops. FIG. 5 is a diagram for explaining a method for determining that there is an obstacle when the vehicle starts when there is no image change between when the vehicle stops and when the vehicle starts in a situation where there is an obstacle when the vehicle stops. FIG. 6 is a diagram for explaining a method for determining that there is an obstacle at the start of the vehicle when there is an image change between the stop of the vehicle and the start of the vehicle in a situation where there is no obstacle when the vehicle is stopped. .

  In the vehicular image recognition apparatus 10 according to the present embodiment, the image processing ECU 14 includes an image picked up by the monocular camera 12, particularly a steering angle sensor 16, a wheel speed sensor 18, and a shift position sensor, particularly during ignition on the host vehicle. The state quantity (specifically, the steering angle, the wheel rotation speed, and the shift shift position) of the host vehicle 20 is acquired (step 100). Each information acquisition such as an image and a state quantity in the image processing ECU 14 is performed at each predetermined sampling time.

  The image processing ECU 14 determines the traveling direction of the host vehicle (specifically, forward or reverse) based on the wheel rotational speed and the shift shift position obtained by the wheel speed sensor 18 and the shift position sensor 20 acquired in step 100. Then, it is determined whether or not the host vehicle is traveling backward (step 102). As a result, when it is determined that the host vehicle is moving backward, a process of detecting an obstacle as a three-dimensional object existing behind the host vehicle using a captured image of the monocular camera 12 (normal three-dimensional object detection process) is performed. Perform (step 104). This three-dimensional object detection process is performed while acquiring parallax of an object between two captured images of the monocular camera 12 obtained in time series at two different points while the host vehicle is moving backward, and acquiring these two captured images. This is performed based on the movement data of the own vehicle. When the image processing ECU 14 completes the three-dimensional object detection process, the image processing ECU 14 next executes the process of step 100 described above.

  If the image processing ECU 14 determines in step 102 that the host vehicle is not moving backward, that is, the host vehicle is moving forward or stopped, the image processing ECU 14 further shifts the wheel speed sensor 18 and the information acquired in step 100. Based on the wheel rotation speed and the shift shift position by the position sensor 20, it is determined whether or not the host vehicle is stopped (step 106). This stop determination is made based on whether the wheel rotation speed is zero, whether the shift shift position is the parking position, or a combination thereof. For example, the determination is affirmative when the speed of the host vehicle is zero. It may be done. As a result, when it is determined that the host vehicle is not stopped, that is, the host vehicle is moving forward, the process of step 100 is executed next.

  On the other hand, if it is determined that the host vehicle is stopped, it is determined whether or not the host vehicle is traveling backward just before the stop (step 108). In this backward determination, immediately before the negative determination is made in step 102 and the positive determination is made in step 106, whether the positive determination is made in step 102 or the three-dimensional object detection process is executed in step 104. It is done based on whether or not.

  If the image processing ECU 14 determines in step 108 that the host vehicle has not moved backward until the vehicle stops, the image processing ECU 14 determines that there is no obstacle existing in the predetermined range behind the host vehicle, and displays the no obstacle information. Store in memory (step 110). In step 102, it is determined that the host vehicle has not moved backward, in step 106, it is determined that the host vehicle has stopped, and in step 108, the host vehicle has moved backward until it stops. If it is determined that the vehicle has not been stopped, it can be determined that the host vehicle was moving forward until the vehicle stopped.Therefore, there should be no obstacles on the movement trajectory that the vehicle passed through until the vehicle stopped. This is because it can be determined that there is no obstacle behind the host vehicle.

  On the other hand, if the image processing ECU 14 determines in step 108 that the host vehicle has moved backward until it stops, the image processing ECU 14 determines that there is an obstacle existing behind the host vehicle by the three-dimensional object detection process during the backward movement. It is determined whether or not it has been done (step 112). This determination is performed based on the result of the three-dimensional object detection process in step 104 described above. As a result, when it is determined that there is an obstacle existing behind the host vehicle during reverse, information on the position of the obstacle existing behind the host vehicle with respect to the host vehicle immediately before the host vehicle stops Is stored in the memory (step 114).

  When the image processing ECU 14 determines in step 110 that there is no obstacle present behind the host vehicle, the image processing ECU 14 determines that it is determined in step 112 that there is no obstacle present behind the host vehicle during reverse. If the obstacle position information is stored in the memory in step 114, the information of the captured image of the monocular camera 12 acquired in step 100 is stored in the memory (step 116).

  Thus, in the present embodiment, whether or not there is an obstacle existing behind the host vehicle by the normal three-dimensional object detection process using the time-series captured images of the monocular camera 12 during the backward movement of the host vehicle. Can be determined. In addition, when the host vehicle is stopped, the information of the captured image of the monocular camera 12 is stored in the memory, the presence / absence information of the obstacle existing behind the host vehicle, and the host vehicle is moving backward immediately before the host vehicle is stopped. When it is determined that there is an obstacle existing behind the host vehicle while the vehicle is leaving, the position information of the obstacle can be further stored in the memory.

  Specifically, when the host vehicle is stopped, if the host vehicle has moved toward the rear of the host vehicle captured by the monocular camera 12 until the vehicle stops (that is, the vehicle has moved backward), The presence / absence information and position information of the obstacle detected based on the time-series change of the captured image of the camera 12 and information of the captured image of the monocular camera 12 when the host vehicle is stopped are stored in memory. On the other hand, when the host vehicle is stopped, if the host vehicle is moving in the direction opposite to the rear of the host vehicle captured by the monocular camera 12 until the vehicle stops (that is, the vehicle is moving forward), the vehicle exists behind the host vehicle. The information indicating that there is no obstacle to be performed and the information of the captured image of the monocular camera 12 when the host vehicle is stopped are stored in memory.

  In this embodiment, the image processing ECU 14 stores in memory the captured image of the monocular camera 12, the presence / absence information of the obstacle, and the position information when the host vehicle stops as described above, and then a predetermined time has elapsed. Or whether or not the engine of the host vehicle is started from a stopped state (step 150). As a result, when a negative determination is made, the process of step 150 is repeatedly executed. On the other hand, when an affirmative determination is made, that is, when it is determined that the predetermined time has elapsed or the own vehicle engine has been started, an image captured by the monocular camera 12 is acquired (step 152). Accordingly, a captured image behind the host vehicle when a predetermined time has elapsed since the host vehicle stopped or when the engine of the host vehicle is started is obtained.

  The image processing ECU 14 compares the current captured image of the monocular camera 12 acquired in step 152 with the captured image of the monocular camera 12 stored in memory in step 116 shown in FIG. A change area detection process is performed to detect a change area between them (step 154). For this change area detection processing, for example, a method described in Japanese Patent Application Laid-Open No. 2009-212453, Japanese Patent No. 383232, or the like may be used. By using this method, it is possible to realize a background difference that is robust against illumination fluctuations and various disturbances in the outdoor environment. The change area to be detected by the change area detection processing is, for example, a predetermined distance (for example, an alarm) from an own vehicle by an obstacle having a size or height that may interfere with traveling of the own vehicle. The distance that the obstacle occupies on the image when it appears in the captured image at a distance (starting the alarm sound by the sound device 24) is set.

  The image processing ECU 14 determines whether or not it is detected that there is a change area between both captured images by the above-described change area detection process (step 156). As a result, if it is detected that there is no change area between the two captured images, then the vehicle will be located behind the vehicle when it stopped last time based on the obstacle information stored in the memory. It is determined whether or not it has been determined that there is an obstacle to be performed (step 158). When there is no change between the current captured image and the captured image when the host vehicle is stopped, it can be determined that there is no change in the state behind the host vehicle from when the host vehicle stops until the current time. Therefore, if it is determined in step 158 that it is determined that there is an obstacle when the host vehicle is stopped, it is determined that there is an obstacle existing behind the host vehicle even in the current state (step 160). On the other hand, if it is determined that there is no obstacle when the host vehicle is stopped, it is determined that there is no obstacle present behind the host vehicle even in the current state (step 162).

  If it is detected in step 156 that there is a change area between the two captured images, the next time the host vehicle stops last time based on the obstacle presence / absence information stored in the memory. It is determined whether or not it has been determined that there is an obstacle existing behind (step 164). When there is a change between the current captured image and the captured image when the host vehicle is stopped, it can be determined that there has been a change in the state behind the host vehicle from when the host vehicle stopped to the present time. In this case, when there is no obstacle when the host vehicle is stopped, there is a possibility that the obstacle has invaded until the present time. On the other hand, when there is an obstacle when the host vehicle is stopped, the obstacle or new It can be determined that the obstacle has moved or entered the captured image, or that the obstacle has gone out of the captured image and no longer exists in the captured image. Therefore, if it is determined in step 164 that it is determined that there is no obstacle when the host vehicle is stopped, it is determined that there is an obstacle existing behind the host vehicle at present (step 166). On the other hand, if it is determined in step 164 that it is determined that there is an obstacle when the host vehicle is stopped, it is determined that it is uncertain whether an obstacle exists behind the host vehicle (step 168).

  The image processing ECU 14 determines whether or not the engine of the host vehicle has been started after determining the current state behind the host vehicle in the above-described step 160, 162, 166, or 168 (step 170). As a result, if it is determined that the engine of the host vehicle has not been started, the processing after step 150 is repeated. On the other hand, if it is determined that the engine of the host vehicle has been started, it is determined or present that there is an obstacle existing behind the host vehicle based on the determination result in step 160, 162, 166, or 168. When it is determined that the answer is indefinite, the display monitor 22 and the alarm sound device 24 are operated so as to perform a display / alarm that alerts the driver of the own vehicle that there is an obstacle or the presence or absence of an obstacle (step 172). ).

  When the image processing ECU 14 starts the operation of the display monitor 22 and the alarm sound device 24 in the above step 172, the state quantity of the host vehicle (specifically, the wheel speed and the shift shift) is determined by the wheel speed sensor 18 and the shift position sensor 20. It is determined whether or not the host vehicle is moving based on (position) (step 174). As a result, when it is determined that the host vehicle is not moving, the processes after step 150 are repeated. On the other hand, if it is determined that the host vehicle is moving, it is not appropriate to perform obstacle detection using the change area detection process based on background difference thereafter. In the same manner, the processing is switched to processing for detecting an obstacle as a three-dimensional object existing behind the host vehicle (normal three-dimensional object detection processing) using the captured image of the monocular camera 12 (step 176). After such switching, an obstacle as a three-dimensional object existing behind the host vehicle is detected by a normal three-dimensional object detection process using parallax between two time-series captured images of the monocular camera 12. The Rukoto.

  As described above, in this embodiment, when a predetermined time has elapsed after the host vehicle stops or when the engine is started, a captured image of the rear of the host vehicle captured by the monocular camera 12 is stored in the memory when the host vehicle stops. Compared with the captured image when the host vehicle is stopped, a change area detection process for detecting a change area between the two is executed, and the change area detection process is repeatedly executed until the movement of the host vehicle is started. To do.

  And, when it is detected by the change area detection process that there is no change area between both captured images, there is no change in the state behind the own vehicle from the time when the own vehicle stops until the present time. Then, the same determination as the obstacle presence / absence determination when the host vehicle is stopped is performed (see FIGS. 4 and 5). On the other hand, when it is detected that there is a change area between both captured images, when it is determined that there is no obstacle when the host vehicle is stopped, it is determined that there is an obstacle as the obstacle presence / absence determination at the present time (FIG. 6). When it is detected that there is a change area between the two captured images and it is determined that there is an obstacle when the host vehicle is stopped, it is determined that the presence or absence of the obstacle at the present time is indefinite. That is, in the present embodiment, the captured image when the host vehicle is stopped and the captured image when the host vehicle is started (precisely, immediately before starting) are compared to detect a change region between the images, and the change region is detected. When there is no obstacle, the obstacle presence / absence determination is performed in the same way as when the vehicle is stopped. On the other hand, if there is a change area, only when it is determined that there is no obstacle when the host vehicle is stopped. Determine that there is an obstacle.

  Therefore, according to the vehicular image recognition apparatus 10 of the present embodiment, the state behind the host vehicle is determined using the captured image of the monocular camera 12 without the movement of the host vehicle when the host vehicle starts after stopping. Can do. In this regard, it is not necessary for the host vehicle to actually start in order to detect the presence of an obstacle existing behind the host vehicle when starting the host vehicle, and time series parallax is added to the captured image of the monocular camera 12. Since it is not necessary to cause it to occur, it is possible to detect the presence or absence of an obstacle existing behind the host vehicle using the captured image of the monocular camera 12 while the host vehicle is stopped. . Furthermore, even if the host vehicle stops when the host vehicle starts and the obstacle existing behind the host vehicle does not move at all, that is, when the host vehicle starts, the host vehicle and the obstacle existing behind the host vehicle are relatively displaced. Even if it does not do, it is possible to implement | achieve the presence detection of the obstruction which exists behind the own vehicle using the captured image of the monocular camera 12. FIG.

  In the above-described embodiment, image processing for storing information on the captured image of the monocular camera 12 when the host vehicle is stopped, presence / absence information of an obstacle existing in the direction in which the monocular camera 12 is captured when the host vehicle is stopped, and position information. The memory in the ECU 14 stores the “storage means” described in the claims, and the image processing ECU 14 executes the processing of steps 110, 114, and 116 in the routine shown in FIG. The image processing ECU 14 executes the processing of steps 160, 162, 166, and 168 in the routine shown in FIG. 3 in the “stationary storage control means”. The ECU 14 performs the normal three-dimensional object detection process in the above step 104 in the “normal obstacle detection means” described in the claims. Which is equivalent.

  By the way, in the above embodiment, the monocular camera 12 for photographing the periphery of the host vehicle is a camera that is attached to the rear part of the vehicle body and shoots an area extending rearward of the host vehicle. It can be applied to a camera that shoots an area that extends to the front of the host vehicle and is attached to the side of the vehicle body and spreads to the side, front side, and rear side. It is good also as applying to the camera which image | photographs an area | region.

  Further, in the above embodiment, after the vehicle is started, the state behind the vehicle (judgment of obstacles) is determined, and then the driver of the vehicle is prompted to be aware of the presence of obstacles or the presence or absence of obstacles. -In order to operate the display monitor 22 and the alarm sound device 24 so as to issue an alarm, it is a condition that the engine is started. After the engine is started, the shift shift position has shifted to the drive range or the reverse range. It may be a condition, or may be a condition that the brake operation is further released.

  Further, in the above embodiment, when it is determined that there is an obstacle existing behind the host vehicle, the display monitor 22 for displaying a warning and an alarm sound are generated to give the driver a warning regarding the obstacle. Although the alarm sound device 24 is activated, the present invention is not limited to this, and a pre-crash safety system for reducing collision damage (for example, instead of or together with these operations) (for example, An automatic braking device, a seat belt restraining device, etc.) may be operated.

  Furthermore, in the above embodiment, the camera 12 for photographing the periphery of the host vehicle is a monocular camera composed of one camera. However, the present invention is not limited to this, and two cameras are used. It may be a stereo camera in which In a modification using such a stereo camera, the host vehicle does not move when the host vehicle starts after stopping, and the host vehicle and the obstacle existing behind the host vehicle are relatively displaced when the host vehicle starts. Even if there is no obstacle, the presence / absence detection of the obstacle can be realized by determining the state behind the host vehicle using the captured image of any one of the stereo cameras. In addition, it is possible to detect the presence or absence of an obstacle existing behind the host vehicle by using a captured image of any one of the stereo cameras without performing a three-dimensional object detection process by the stereo camera.

DESCRIPTION OF SYMBOLS 10 Image recognition apparatus for vehicles 12 Imaging device (monocular camera)
14 Electronic control unit (image processing ECU)
34 CPU

Claims (8)

A camera that captures the area around the vehicle,
A stop time storage control means for storing in the storage means information indicating a stop time state around the own vehicle when the own vehicle is stopped;
State determining means for determining a state around the own vehicle based on a captured image of the camera and the stop state stored in the storage means when the host vehicle starts;
A vehicle image recognition apparatus comprising:   The vehicle image recognition device according to claim 1, wherein the stop state includes an image captured by the camera when the host vehicle is stopped.   The vehicular image recognition apparatus according to claim 2, wherein the stop state further includes the presence / absence and position of an obstacle present in a direction in which the camera takes a picture when the host vehicle stops.   When the host vehicle has traveled in a direction opposite to the direction in which the camera takes a picture until the host vehicle stops, information indicating that the obstacle is absent when the host vehicle stops and the host vehicle stop 4. The vehicular image recognition apparatus according to claim 3, wherein image information of the camera at the time is stored in the storage means.   The stop time storage control means is based on the time-series change of the captured image of the camera until the host vehicle stops when the host vehicle is traveling in the direction of shooting by the time the host vehicle stops. When the presence / absence information of the detected obstacle and the presence of the obstacle are detected, the storage means stores the position information of the obstacle and the imaged image information of the camera when the host vehicle is stopped. The vehicular image recognition apparatus according to claim 3 or 4, characterized by the above-mentioned.   6. The vehicle according to claim 5, further comprising a normal obstacle detection unit that detects the presence / absence and position of the obstacle based on a time-series change of a captured image of the camera until the host vehicle stops. Image recognition device.   The state determining means may be arranged when the vehicle starts, and when there is no change more than a predetermined value between the captured image of the camera and the captured image of the camera stored in the storage means when the vehicle is stopped. The vehicle state according to any one of claims 3 to 6, wherein the state is determined to be in the same state as the presence or absence of the obstacle when the host vehicle is stopped, the information being stored in the storage means. Image recognition device. The state determining means may store the storage means when there is a predetermined change or more between the captured image of the camera when the host vehicle starts and the captured image of the camera stored when information is stored in the storage means. 8. When information indicating that there is no obstacle when the host vehicle is stopped is stored in the vehicle, it is determined that the obstacle exists as a state around the host vehicle. The image recognition apparatus for vehicles as described.

Images