JP2009132259A - Vehicle surrounding-monitoring device - Google Patents

Abstract

<P>PROBLEM TO BE SOLVED: To provide a vehicle surrounding-monitoring device capable of enhancing effect of informing by reducing unnecessary informing to a driver. <P>SOLUTION: The vehicle surrounding monitoring device 1 includes a detection area setting part 13, an obstacle detection part 18, and an informing determination part 15. The the detection area setting part 13 sets three detection areas different in size according to the distance from a vehicle on an expected course of the vehicle in a camera image based on steering information of the vehicle and information of camera installing conditions. The obstacle detection part 18 provides image processing for each of the three detection areas and detects an obstacle in the detection area. The informing determination part 15 determines whether the detection informing of the obstacle is performed or not based on the distance from the vehicle of the detection area in which the obstacle is detected. <P>COPYRIGHT: (C)2009,JPO&INPIT

Description

  The present invention relates to a vehicle periphery monitoring device that detects an obstacle around a vehicle using a camera mounted on the vehicle.

2. Description of the Related Art Conventionally, there has been known an object detection device that captures a camera image around a vehicle using a camera mounted on the vehicle, performs image processing on the camera image, and detects an obstacle around the vehicle (for example, Patent Document 1). reference). In this conventional object detection apparatus, an obstacle in the comparison image is detected by performing a difference process comparing a reference image captured in advance (for example, immediately after stopping) and a comparison image sequentially captured thereafter. Further, by comparing the two comparison images (the previous comparison image and the current comparison image) with each other, it is determined whether the obstacle is a stationary object or a moving object.
JP 2000-182027 A (page 4-10, FIG. 6)

  However, in the conventional object detection device, all obstacles in the camera image are detected regardless of the distance from the vehicle to the obstacle. If all the obstacles detected in this way are uniformly notified to the driver, the driver is also notified of obstacles far from the vehicle that do not affect safe driving. When notifications are made even for obstacles that do not need to be reported in this way, the notification effect (attention effect) on the obstacles that really need to be notified (for example, obstacles near the vehicle that are highly dangerous) May become diluted.

  Further, in the conventional object detection device, it is necessary to take a reference image while the vehicle is stopped, and no consideration is given to the case where the vehicle is running. Further, although it can be determined whether the detected obstacle is a stationary object or a moving object, it is possible to determine what the obstacle is (for example, a person or a vehicle). It may not be possible to erroneously detect an obstacle.

  The present invention has been made in view of the above-described problems, and an object of the present invention is to provide a vehicle periphery monitoring device that can reduce the notification unnecessary for the driver and enhance the effect of the notification.

  A vehicle periphery monitoring device according to the present invention is a vehicle periphery monitoring device that detects an obstacle around a vehicle using a camera image taken by a camera mounted on the vehicle, the steering angle information of the vehicle steering and the camera Detection area setting means for setting a plurality of detection areas having different sizes in accordance with the distance from the vehicle on the expected course of the vehicle in the camera image, and the plurality of detection areas And detecting obstacles based on the distance from the vehicle to the detection area where the obstacle is detected, and obstacle detection means for detecting an obstacle in the detection area. Notification determining means for determining whether to perform or not.

  Thus, a plurality of detection areas are set for each distance from the vehicle, and obstacles are detected for each detection area. Then, it is determined whether or not to perform the detection notification based on the distance from the vehicle in the detection area where the obstacle is detected. For example, the detection notification of the obstacle near the vehicle is performed, and the detection notification of the obstacle far away from the vehicle is not performed. Thereby, notifications unnecessary for the driver can be reduced. Further, since it is not necessary to take a reference image while the vehicle is stopped as in the prior art, it is possible to detect and notify an obstacle even when the vehicle is traveling.

  Further, in the vehicle periphery monitoring device of the present invention, the detection area setting means has a virtual cross-sectional area formed by cutting a virtual three-dimensional zone along the predicted course of the vehicle according to the distance from the vehicle, The detection area is set as the detection area, and when the distance of the detection area from the vehicle is equal to or less than a predetermined threshold distance, the notification determination unit performs detection notification of the obstacle, and the vehicle in the detection area If the distance from the vehicle is larger than the threshold distance, it may be determined not to notify the obstacle detection.

  Thereby, the virtual cross-sectional area obtained by cutting the virtual three-dimensional zone along the predicted course of the vehicle for each distance from the vehicle is set as the detection region. In this way, the detection region is appropriately set by considering the three-dimensional element. For example, when a detection area is set in a two-dimensional manner as in the past, only a part of a vertically long obstacle (for example, only a pedestrian's foot) enters the detection area. It may be erroneously detected (detected as an obstacle other than a pedestrian). On the other hand, in the present invention, since the detection area is appropriately set in consideration of a three-dimensional element, even if it is a vertically long obstacle, the entire area enters the detection area, thereby reducing false detection of the obstacle. be able to. Further, in this case, detection notification is performed for obstacles closer than a predetermined threshold distance, and detection notification is not performed for obstacles farther than the threshold distance. Thereby, notifications unnecessary for the driver can be reduced.

  In the vehicle periphery monitoring device according to the present invention, the obstacle detection unit may be a storage unit that stores model pattern data of the obstacle, and a detection area where the obstacle is detected is detected according to a distance from the vehicle. Pattern matching means for adjusting a pattern size and performing pattern matching processing for determining whether or not the pattern detected in the detection area matches the model pattern of the obstacle, and the notification determining means Depending on the result of the pattern matching process, it may be determined whether or not the obstacle detection notification is performed.

  By performing the pattern matching process in this way, it is possible to determine what the detected obstacle is (for example, whether it is a person or a vehicle). And according to the result of the pattern matching process, the obstacle detection notification is appropriately performed. For example, if the detected obstacle is a “child” pedestrian as a result of the pattern matching process, the detection is performed because the danger level is high, and the detected obstacle is an “adult” pedestrian. Is not detected because the degree of danger is relatively low.

  In the vehicle periphery monitoring device of the present invention, a plurality of cameras are mounted on the vehicle, and the camera used for detecting the obstacle from the camera images of the plurality of cameras based on shift gear information of the vehicle. You may provide the image selection means to select an image.

  Thereby, the camera image used for the detection of an obstacle is appropriately switched according to the shift gear operation. For example, when the shift gear is set to “R”, the camera image is switched to the rear of the vehicle, and an obstacle on the path behind the vehicle can be detected. Further, when the shift gear is set to “D”, the camera image is switched to the front of the vehicle, and an obstacle on the course ahead of the vehicle can be detected.

  In the vehicle periphery monitoring device of the present invention, the notification determining unit may include a threshold distance changing unit that changes the threshold distance based on speed information of the vehicle.

  Thereby, the reference | standard which performs detection notification of an obstruction is changed appropriately according to the running speed of a vehicle. For example, when the traveling speed of the vehicle is low, the detection notification of the obstacle near the vehicle is performed, and the detection notification of the obstacle far away from the vehicle is not performed because the risk is low. Further, when the traveling speed of the vehicle is high, the detection notification is performed because the risk is high even for an obstacle far from the vehicle.

  The vehicle periphery monitoring method of the present invention is a vehicle periphery monitoring method for detecting obstacles around the vehicle using a camera image taken by a camera mounted on the vehicle, the steering angle information of the vehicle steering and the camera On the basis of the installation condition information, a plurality of detection areas having different sizes according to the distance from the vehicle are set on the expected course of the vehicle in the camera image, and image processing is performed on each of the plurality of detection areas. The obstacle is detected in the detection area, and it is determined whether or not to detect the obstacle based on the distance from the vehicle in the detection area where the obstacle is detected.

  According to this method, a plurality of detection areas are set for each distance from the vehicle, and obstacles are detected for each detection area. Then, it is determined whether or not to perform the detection notification based on the distance from the vehicle in the detection area where the obstacle is detected. For example, the detection notification of the obstacle near the vehicle is performed, and the detection notification of the obstacle far away from the vehicle is not performed. Thereby, notifications unnecessary for the driver can be reduced. Further, since it is not necessary to take a reference image while the vehicle is stopped as in the prior art, it is possible to detect and notify an obstacle even when the vehicle is traveling.

  The vehicle periphery monitoring program according to the present invention is a vehicle periphery monitoring program for detecting obstacles around the vehicle using a camera image taken by a camera mounted on the vehicle, and the steering angle information of the steering of the vehicle is stored in a computer. A plurality of detection areas having different sizes according to the distance from the vehicle on the expected course of the vehicle in the camera image based on the installation condition information of the camera The obstacle is detected based on an obstacle detection process for detecting an obstacle in the detection area by performing image processing on each of the detection areas, and a distance from the vehicle to the detection area where the obstacle is detected. And a notification determination process for determining whether or not to perform the detection notification.

  According to this program, a plurality of detection areas are set for each distance from the vehicle, and obstacles are detected for each detection area. Then, it is determined whether or not to perform the detection notification based on the distance from the vehicle in the detection area where the obstacle is detected. For example, the detection notification of the obstacle near the vehicle is performed, and the detection notification of the obstacle far away from the vehicle is not performed. Thereby, notifications unnecessary for the driver can be reduced. Further, since it is not necessary to take a reference image while the vehicle is stopped as in the prior art, it is possible to detect and notify an obstacle even when the vehicle is traveling.

  According to the vehicle periphery monitoring device of the present invention, it is possible to reduce the notification unnecessary for the driver and enhance the notification effect.

  Hereinafter, a vehicle periphery monitoring device according to an embodiment of the present invention will be described with reference to the drawings. In the present embodiment, a case where the present invention is used for a reverse parking assist device for a vehicle is illustrated. The reverse parking assist device detects an obstacle behind the vehicle and informs the driver when the vehicle is parked backward. Such a reverse parking support function is realized by a program stored in the HDD or memory of the apparatus.

  The vehicle periphery monitoring apparatus of this Embodiment is demonstrated using FIGS. 1-10. Here, first, the configuration of the vehicle periphery monitoring device will be described. FIG. 1 is a block diagram showing a main configuration of a vehicle periphery monitoring apparatus.

  As shown in FIG. 1, the vehicle periphery monitoring device 1 includes a navigation control unit 2 and an image processing control unit 3 configured by a CPU, a memory, and the like. The navigation control unit 2 has a navigation function for performing voice guidance and image guidance for backward parking assistance and the like. The image processing control unit 3 has a function of performing image processing for backward parking assistance and the like. Detailed configurations of the navigation control unit 2 and the image processing control unit 3 will be described later.

  A plurality of cameras 4 are mounted on the vehicle. For example, a front camera that captures the front of the vehicle, a side camera that captures the side of the vehicle, a rear camera that captures the rear of the vehicle, and the like are mounted on the vehicle. In FIG. 1, for convenience of explanation, a plurality of cameras 4 are not shown, and only one camera 4 (for example, a rear camera) is shown. And the camera image image | photographed with the camera 4 is input into the navigation control unit 2 and the image processing control unit 3, respectively.

  Further, the vehicle includes a shift gear switch 5 for detecting a shift gear position (for example, “1, 2, D, N, R, P”, etc.) and a steering angle sensor 6 for detecting a steering angle (cutting angle) of the steering. And a speed sensor 7 for detecting the traveling speed of the vehicle. The navigation control unit 2 receives shift gear information indicating the shift gear position from the shift gear switch 5, receives steering angle information indicating the steering angle of the steering from the steering angle sensor 6, and determines the traveling speed of the vehicle from the speed sensor 7. The indicated speed information is input.

  As shown in FIG. 1, the vehicle periphery monitoring device 1 includes a monitor 8 and a speaker 9. The monitor 8 receives image data from the navigation control unit 2 and displays image guidance for backward parking assistance and the like. In addition, voice data is sent from the navigation control unit 2 to the speaker 9 to provide voice guidance for backward parking assistance and the like. In the present embodiment, detection notification when an obstacle around the vehicle is detected is performed by sound from the speaker 9.

  FIG. 2 is a block diagram showing the configuration of the navigation control unit 2. As shown in FIG. 2, the navigation control unit 2 includes an image acquisition unit 10 that acquires an image from the camera 4. The navigation control unit 2 also includes an information acquisition unit 11 that acquires shift gear information, steering angle information, and speed information from the shift gear switch 5, the steering angle sensor 6, and the speed sensor 7. The navigation control unit 2 includes an image selection unit 12 that selects a camera image for obstacle detection from the camera images of the plurality of cameras 4 based on the shift gear information. For example, when the shift gear position is “R (reverse)”, the camera image of the rear camera 4 is selected.

  As shown in FIG. 2, the navigation control unit 2 includes a detection area setting unit 13 that sets a detection area for detecting an obstacle around the vehicle. The navigation control unit 2 includes a detection area database 14, and the detection area database 14 stores detection area setting data for setting a detection area in a camera image. This detection area setting data includes detection areas having different sizes according to the distance from the vehicle based on steering angle information (for example, 0 degrees, 60 degrees, 120 degrees,...) And detection area setting data. This is data for arranging on the expected course of the vehicle in the camera image. Note that the installation condition information of the camera 4 includes specification conditions of the camera 4 (for example, angle of view, resolution, etc.).

  For example, the detection area setting data includes the position of the detection area (for example, the coordinates of the corners of the square frame) and the size (for example, the vertical and horizontal lengths of the square frame) for each predicted course of the vehicle and the distance from the vehicle. And is stored in advance in the detection area database 14 at the time of product shipment. The position of the detection area is arranged at a position shifted in the left-right direction according to the steering angle of the steering. In addition, the size of the detection area is set to be smaller as the detection area is for detecting an obstacle that is far from the vehicle. Specifically, a virtual three-dimensional zone along the expected course of the vehicle is assumed in the camera image, and the virtual three-dimensional zone is determined for each distance (for example, 0.5 m, 1.5 m, 3.0 m) from the vehicle. Three virtual cross-sectional areas formed when cutting are set as detection areas (see FIGS. 7 to 9). In the figure, a detection area with a distance of 0.5 m from the vehicle is indicated by a broken line, a detection area of 1.5 m is indicated by a one-dot chain line, and a detection area of 3.0 m is indicated by a two-dot chain line. ing.

  The navigation control unit 2 includes a notification determining unit 15 that determines whether or not to perform obstacle detection notification. The notification determination unit 15 determines whether to perform detection notification based on the distance from the vehicle in the detection area when an obstacle is detected by pattern matching described later. Specifically, the detection notification is performed when the distance from the vehicle in the detection area is 2.0 m or less, and the detection notification is not performed when the distance from the vehicle in the detection area is greater than 2.0 m. . That is, detection notification is performed when the distance from the vehicle is 0.5 m and 1.5 m, and detection notification is not performed when the distance from the vehicle is 3.0 m. In this case, it can be said that the distance of 2.0 m from the vehicle is a threshold distance that serves as a reference for whether or not detection notification is performed.

  The notification determining unit 15 includes a threshold distance changing unit 16 that changes the threshold distance based on vehicle speed information. For example, the threshold distance changing unit 16 sets the threshold distance to 2.0 m when the vehicle traveling speed is 10 km / h or less, and when the vehicle traveling speed is 10 km / h or more, Set the threshold distance to 5.0 m. As a result, when the traveling speed of the vehicle is 10 km / h or less, the obstacle is detected and notified at the positions of 0.5 m and 1.5 m from the vehicle, and the obstacle at the position of 3.0 m from the vehicle. No detection notification is performed. On the other hand, when the traveling speed of the vehicle is 10 km / h or more, even if the obstacle is located at a position of 3.0 m from the vehicle, the detection notification is performed.

  FIG. 3 is a block diagram showing a configuration of the image processing control unit 3. As shown in FIG. 3, the image processing control unit 3 includes an image acquisition unit 17 that acquires an image from the camera 4 and an obstacle detection unit 18 that detects an obstacle in the detection region. The obstacle detection unit 18 is a pattern matching database 19 that stores data of an obstacle model pattern, and a pattern matching process that determines whether the pattern detected in the detection region matches the obstacle model pattern. A pattern matching unit 20 is provided.

  For example, the pattern matching unit 20 performs pattern matching processing using a machine learning method such as CNN (Convolutional Neural Network). The obstacle model patterns include model patterns of pedestrians (adult pedestrians, child pedestrians, etc.), vehicles (automobiles, motorcycles, bicycles, etc.). Such obstacle model pattern data is previously stored in the pattern matching database 19 by machine learning. The contents of the pattern matching process will be described later.

  The backward parking assist function of the vehicle periphery monitoring device 1 configured as described above will be described with reference to FIGS. FIG. 4 is an explanatory diagram showing a situation where a camera 4 captures a pedestrian behind the vehicle when parked backward, and FIG. 5 is an explanatory diagram of a camera image captured at that time.

  As shown in FIG. 4, in this example, the vehicle goes straight toward the parking space behind, and pedestrians exist at distances of 0.5 m, 1.5 m, and 3.0 m from the vehicle. In this case, as shown in FIG. 5, the camera image of the rear camera 4 is provided with obstacle detection areas of 0.5 m, 1.5 m, and 3.0 m, and three pedestrians are located in the respective detection areas. Taken on.

  FIGS. 6A to 6C are explanatory diagrams showing how pattern matching processing is performed. FIG. 6A is an explanatory diagram of pattern matching of an obstacle with a distance of 0.5 m from the vehicle. In this case, first, a detection area (dashed square frame in FIG. 5) having a distance of 0.5 m is reduced (or enlarged) to a pattern matching size (predetermined size). Then, it is determined whether or not the reduced (or enlarged) pattern in the detection area matches the model pattern in the pattern matching database 19. In this example, as shown with diagonal lines in the figure, the pattern of the left pedestrian (pedestrian at a position of 0.5 m) in the camera image matches the model pattern for pattern matching. In this way, a pedestrian at a position of 0.5 m is detected as an obstacle.

  FIG. 6B is an explanatory diagram of pattern matching of an obstacle with a distance of 1.5 m from the vehicle. In this case, a detection area (a dashed-dotted square frame in FIG. 5) having a distance of 1.5 m is reduced (or enlarged) to a pattern matching size (predetermined size). Then, it is determined whether or not the reduced (or enlarged) pattern in the detection area matches the model pattern in the pattern matching database 19. In this example, the pattern of the central pedestrian (pedestrian at a position of 1.5 m) in the camera image matches the model pattern for pattern matching, as indicated by using diagonal lines in the figure. In this way, a pedestrian at a position of 1.5 m is detected as an obstacle.

  FIG. 6C is an explanatory diagram of pattern matching of an obstacle with a distance of 3.0 m from the vehicle. In this case, the detection area (distance chain square frame in FIG. 5) having a distance of 3.0 m is reduced (or enlarged) to a pattern matching size (predetermined size). Then, it is determined whether or not the reduced (or enlarged) pattern in the detection area matches the model pattern in the pattern matching database 19. In this example, as shown with diagonal lines in the figure, the pattern of the right pedestrian (pedestrian at a position of 3.0 m) in the camera image matches the model pattern for pattern matching. In this way, a pedestrian at a position of 3.0 m is detected as an obstacle.

  7 to 9 are diagrams showing how the arrangement of the detection regions changes when the steering is turned off. FIG. 7 is an explanatory view showing the arrangement of the detection areas when the vehicle is parked in a straight line as in FIG. As shown in FIG. 7, when the steering is made straight, a virtual cross-sectional area obtained by cutting a virtual three-dimensional zone along an expected course when going backward while going straight is set as a detection region.

  In the example of FIG. 7, a pedestrian (a distance of 3.0 m from the vehicle) is shown on the right side of the center of the camera screen. This pedestrian is included in a detection area with a distance of 3.0 m and is detected as an obstacle with a distance of 3.0 m because it matches the model pattern in the pattern matching process. Although this pedestrian is also included in the detection area at a distance of 0.5 m or 1.5 m, the size of the model pattern does not match in the pattern matching process, so the distance of 0.5 m or 1.5 m It is not detected as an obstacle.

  FIG. 8 is an explanatory diagram showing the arrangement of detection areas when the steering is turned to the left. As shown in FIG. 8, when the steering is turned to the left, a virtual cross-sectional area obtained by cutting a virtual three-dimensional zone along an expected course that turns backward while turning right is set as a detection region.

  In the example of FIG. 8 as well, a pedestrian (a distance of 3.0 m from the vehicle) is shown at the same position as in FIG. This pedestrian is included in a detection area with a distance of 3.0 m and is detected as an obstacle with a distance of 3.0 m because it matches the model pattern in the pattern matching process. In FIG. 8, since the detection area is arranged on the expected course that turns backward while turning to the right, a pedestrian on the expected course can be detected at the center of the detection area as compared with FIG. Although this pedestrian is also included in the detection area at a distance of 0.5 m or 1.5 m, the size of the model pattern does not match in the pattern matching process, so the distance of 0.5 m or 1.5 m It is not detected as an obstacle.

  FIG. 9 is an explanatory diagram showing the arrangement of detection areas when the steering is turned to the right. As shown in FIG. 9, when the steering is turned to the right, a virtual cross-sectional area obtained by cutting a virtual three-dimensional zone along an expected course turning backward while turning left is set as a detection region.

  Also in the example of FIG. 9, a pedestrian (a distance of 3.0 m from the vehicle) is shown at the same position as in FIG. In this case, since the detection area is arranged on the expected course of turning backward while turning left, the pedestrian is not included in the detection area of the distance of 3.0 m and is not detected as an obstacle of the distance of 3.0 m. In addition, although this pedestrian is included in a detection area with a distance of 0.5 m or 1.5 m, the size of the model pattern does not match in the pattern matching process, so the distance of 0.5 m or 1.5 m It is not detected as an obstacle.

  FIG. 10 is a flowchart showing a flow of operations of the reverse parking support function of the vehicle periphery monitoring device 1. As shown in FIG. 10, when the reverse parking support function is turned on, the navigation control unit 2 acquires vehicle shift gear information from the shift gear switch 5 (S1). Then, it is determined whether or not the position of the shift gear is “R (reverse)” (S2).

  When the position of the shift gear is “R”, the navigation control unit 2 acquires a camera image of the rear camera 4 of the vehicle (S3). Further, the navigation control unit 2 acquires the steering angle information of the steering from the steering angle sensor 6 (S4), and determines the expected course of the vehicle (S5). Then, the navigation control unit 2 sets three detection areas on the predicted course of the vehicle in the camera image based on the acquired steering angle information and the detection area setting data read from the detection area database 14 (S6). .

  Then, the image processing control unit 3 performs pattern matching processing for each of the three detection areas (S7), and determines whether there is an obstacle in the detection area (S8). When an obstacle is detected in a certain detection area, it is determined whether or not the distance from the vehicle in the detection area is equal to or less than a threshold distance (S9). When the distance from the vehicle in the detection area where the obstacle is detected is equal to or less than the threshold distance, a voice notification to the driver is given from the speaker 9 (S10).

  According to the vehicle periphery monitoring apparatus 1 of the present embodiment as described above, notifications unnecessary for the driver can be reduced and the notification effect can be enhanced. Further, it is possible to detect and notify an obstacle even while the vehicle is traveling.

  That is, in the present embodiment, three detection areas are set for each distance from the vehicle (for example, 0.5 m, 1.5 m, and 3.0 m), and obstacle detection is performed for each detection area. Then, it is determined whether or not to perform the detection notification based on the distance from the vehicle in the detection area where the obstacle is detected. For example, an obstacle near the vehicle (for example, an obstacle at a position of 0.5 m and 1.5 m from the vehicle) is detected and notified, and an obstacle far from the vehicle (for example, 3.0 m from the vehicle) Detection notification of the obstacle at the position is not performed. Thereby, notifications unnecessary for the driver can be reduced. Further, since it is not necessary to take a reference image while the vehicle is stopped as in the prior art, it is possible to detect and notify an obstacle even when the vehicle is traveling.

  In this case, a virtual cross-sectional area obtained by cutting a virtual three-dimensional zone along the expected course of the vehicle for each distance from the vehicle is set as a detection region. In this way, the detection region is appropriately set by considering the three-dimensional element. For example, when a detection area is set in a two-dimensional manner as in the past, only a part of a vertically long obstacle (for example, only a pedestrian's foot) enters the detection area. It may be erroneously detected (detected as an obstacle other than a pedestrian). On the other hand, in the present embodiment, the detection area is appropriately set in consideration of the three-dimensional elements, so even if the obstacle is a vertically long obstacle, the whole enters the detection area, and erroneous detection of the obstacle is performed. Can be reduced. Further, in this case, detection notification is performed for an obstacle near a predetermined threshold distance (for example, 2 m), and detection notification is not performed for an obstacle far from the threshold distance. Thereby, notifications unnecessary for the driver can be reduced.

  In the present embodiment, it is possible to determine what the detected obstacle is (for example, whether it is a person or a vehicle) by performing the pattern matching process. And according to the result of the pattern matching process, the obstacle detection notification is appropriately performed. For example, if the detected obstacle is a “child” pedestrian as a result of the pattern matching process, the detection is performed because the danger level is high, and the detected obstacle is an “adult” pedestrian. Is not detected because the degree of danger is relatively low.

  For example, in a conventional object detection device, even when a tree branch or shadow of an implanted plant enters the detection area, it is erroneously detected as an obstacle (a thing that is not an obstacle is mistakenly regarded as an obstacle) May be detected). On the other hand, in the present embodiment, pattern matching processing is performed to detect obstacles, so that objects that do not match the obstacle model pattern (such as branches and shadows of trees or the like) are detected as obstacles. Therefore, it is possible to reduce false detection of obstacles.

  In addition, in the conventional object detection device, even if it should be detected as an obstacle, if the obstacle is included in the reference image, it is treated as a background in the difference processing with the comparison image, May not be detected. On the other hand, in this embodiment, since the obstacle is detected by performing the pattern matching process, the obstacle that should be detected can be appropriately detected.

  Moreover, in this Embodiment, the camera image used for an obstacle detection is switched appropriately according to shift gear operation. For example, when the shift gear is set to “R”, the camera image is switched to the rear of the vehicle, and an obstacle on the path behind the vehicle can be detected. Further, when the shift gear is set to “D”, the camera image is switched to the front of the vehicle, and an obstacle on the course ahead of the vehicle can be detected.

  Moreover, in this Embodiment, the reference | standard which performs detection notification of an obstruction is changed appropriately according to the running speed of a vehicle. For example, when the traveling speed of the vehicle is low (for example, 10 km / h or less), the detection and notification of an obstacle near the vehicle (for example, an obstacle at positions 0.5 m and 1.5 m from the vehicle) is performed. However, detection and notification of an obstacle far from the vehicle (for example, an obstacle at a position of 3.0 m from the vehicle) is not performed because the risk is low. Further, when the traveling speed of the vehicle is high (for example, 10 km / h or more), even if the obstacle is far away from the vehicle, the danger is high, and thus the detection notification is performed.

  The embodiments of the present invention have been described above by way of example, but the scope of the present invention is not limited to these embodiments, and can be changed or modified according to the purpose within the scope of the claims. is there.

  In the above description, the example in which the image selection unit 12 selects the front and rear camera images of the vehicle based on the shift gear information has been described, but the scope of the present invention is not limited to this. For example, the image selection unit 12 may select a camera image on the side of the vehicle as a camera image used for detecting an obstacle based on the blinker information of the vehicle.

  Moreover, although the detection area setting part 13 demonstrated the example which arrange | positions a detection area on an estimated course based on steering angle information in the above description, the scope of the present invention is not limited to this. For example, the detection area setting unit 13 may arrange the detection area on the expected course based on the blinker information of the vehicle.

  Moreover, although the above description demonstrated the example which performs the detection notification of an obstruction with the audio | voice using the speaker 9, the scope of the present invention is not limited to this. For example, the obstacle detection notification may be performed by highlighting the obstacle detection position on the image of the monitor 8 as well as the sound from the speaker 9.

  As described above, the vehicle periphery monitoring device according to the present invention has an effect that the notification effect unnecessary for the driver can be reduced and the effect of the notification can be enhanced. is there.

It is a block diagram which shows the structure of the vehicle periphery monitoring apparatus in this Embodiment. It is a block diagram which shows the structure of a navigation control unit. It is a block diagram which shows the structure of an image processing control unit. It is explanatory drawing which shows a mode that a pedestrian behind a vehicle is image | photographed when carrying out reverse parking. It is explanatory drawing of the camera image when carrying out reverse parking. (A) It is explanatory drawing of the pattern matching using the detection area | region for obstacles of the distance of 0.5 m from a vehicle. (B) It is explanatory drawing of the pattern matching using the detection area | region for obstacles of distance 1.5m from a vehicle. (C) It is explanatory drawing of the pattern matching using the detection area | region for obstacles of distance 3.0m from a vehicle. It is explanatory drawing which shows arrangement | positioning of the detection area | region when carrying out receding parking in a straight line. It is explanatory drawing which shows arrangement | positioning of a detection area | region when turning a steering to the left and carrying out reverse parking. It is explanatory drawing which shows arrangement | positioning of the detection area | region when turning a steering to the right and carrying out reverse parking. It is a flowchart for demonstrating the flow of operation | movement of a vehicle periphery monitoring apparatus.

Explanation of symbols

DESCRIPTION OF SYMBOLS 1 Vehicle periphery monitoring apparatus 2 Navigation control unit 3 Image processing control unit 4 Camera 5 Shift gear switch 6 Steering angle sensor 7 Speed sensor 8 Monitor 9 Speaker 10 Image acquisition part 11 Information acquisition part 12 Image selection part 13 Detection area setting part 14 Detection area Database 15 Notification determining unit 16 Threshold distance changing unit 17 Image acquiring unit 18 Obstacle detecting unit 19 Pattern matching database 20 Pattern matching unit

Claims (7)

A vehicle periphery monitoring device that detects obstacles around a vehicle using a camera image taken by a camera mounted on the vehicle,
Based on the steering angle information of the steering of the vehicle and the installation condition information of the camera, a plurality of detection areas having different sizes according to the distance from the vehicle are set on the expected course of the vehicle in the camera image. Detection area setting means;
Obstacle detection means for performing an image process on each of the plurality of detection areas and detecting an obstacle in the detection area;
Notification determining means for determining whether to perform detection notification of the obstacle based on the distance from the vehicle of the detection area where the obstacle is detected;
A vehicle periphery monitoring device comprising: The detection area setting means includes:
A virtual cross-sectional area formed by cutting a virtual three-dimensional zone along the predicted course of the vehicle according to the distance from the vehicle, is set as the detection region;
The notification determining means includes
When the distance of the detection area from the vehicle is equal to or less than a predetermined threshold distance, the obstacle is detected and notified, and when the distance of the detection area from the vehicle is larger than the threshold distance The vehicle periphery monitoring device according to claim 1, wherein the vehicle periphery monitoring device is determined not to perform detection and notification of the obstacle. The obstacle detection means includes
Storage means for storing the model pattern data of the obstacle;
A pattern for adjusting whether the pattern detected in the detection area matches the model pattern of the obstacle by adjusting the pattern size according to the distance from the vehicle in the detection area where the obstacle is detected Pattern matching means for performing matching processing;
With
The notification determining means includes
The vehicle periphery monitoring apparatus according to claim 1 or 2, wherein whether or not to perform detection notification of the obstacle is determined according to a result of the pattern matching processing. The vehicle is equipped with a plurality of cameras,
The image selection means which selects the camera image used for the detection of the said obstruction from the camera image of the said several camera based on the shift gear information of the said vehicle, The any one of Claims 1-3 characterized by the above-mentioned. The vehicle periphery monitoring device described in 1.   5. The vehicle periphery monitoring according to claim 2, wherein the notification determining unit includes a threshold distance changing unit that changes the threshold distance based on speed information of the vehicle. apparatus. A vehicle periphery monitoring method for detecting obstacles around a vehicle using a camera image taken by a camera mounted on the vehicle,
Based on the steering angle information of the vehicle steering and the installation condition information of the camera, a plurality of detection areas having different sizes according to the distance from the vehicle are set on the expected course of the vehicle in the camera image. ,
Image processing is performed on each of the plurality of detection areas to detect obstacles in the detection areas,
A vehicle periphery monitoring method, comprising: determining whether or not to perform the obstacle detection detection based on a distance from the vehicle in a detection area where the obstacle is detected. A vehicle periphery monitoring program for detecting obstacles around a vehicle using a camera image taken by a camera mounted on the vehicle,
On the computer,
Based on the steering angle information of the steering of the vehicle and the installation condition information of the camera, a plurality of detection areas having different sizes according to the distance from the vehicle are set on the expected course of the vehicle in the camera image. Detection area setting process;
Obstacle detection processing for performing an image process on each of the plurality of detection areas to detect an obstacle in the detection area;
A notification determination process for determining whether to perform detection notification of the obstacle, based on a distance from the vehicle in a detection region where the obstacle is detected;
The vehicle periphery monitoring program characterized by performing this.

Images