JP2011248640A - Vehicle periphery monitoring device - Google Patents

Abstract

PROBLEM TO BE SOLVED: To notify a driver of objects around a vehicle including pedestrians by detecting the objects from images acquired by imaging means that is installed to the vehicle.SOLUTION: The vehicle periphery monitoring device comprises: imaging means that is installed to a vehicle for acquiring images around the vehicle; object identification means for identifying the objects around the vehicle for the images acquired by the imaging means; motion vector calculation means for calculating the motion vectors of the objects; type determination means for determining object types; and notification means for notifying the driver of the objects when the object is a specified target and the vehicle and the object are placed in specified relation. The type determination means performs: (i) first type determination whose accuracy is high at heavy-operation loads when the objects exist in a first area which the object reaches in time either shorter than or equal to a specified time; and (ii) second type determination whose accuracy is low at operation loads lighter than those of the first type determination when the object exists in a second area which the object reaches in time longer than the specified time.

Description

  The present invention relates to a vehicle periphery monitoring device, and more specifically, for detecting an object around a vehicle including a pedestrian and notifying a driver from an image obtained by an imaging unit mounted on the vehicle. The present invention relates to a vehicle periphery monitoring device.

  Patent Document 1 discloses a periphery monitoring device that monitors the periphery of a vehicle in order to detect an object that may collide with the vehicle from an image obtained by an imaging unit mounted on the vehicle.

Japanese Patent No. 3515926

  In the periphery monitoring device described in Patent Literature 1, when an object such as a pedestrian is detected, a frame is added to the detection object on the display screen so that the driver can easily recognize it. The presence of the frame is emphasized by changing the color of the frame or generating a warning sound.

  However, in the conventional peripheral monitoring device such as Patent Document 1, when the frame color is changed according to the detected position of the object or a warning sound is generated, the detection state of the object, that is, the past at a predetermined timing The same image processing is carried out regardless of whether a frame has already been added or a warning sound has been generated. For this reason, particularly in a situation where a plurality of processes must be performed in parallel, there is a risk that the amount of calculation required for the processes will be compressed.

  Therefore, the present invention reduces or eliminates this problem of the prior art, that is, optimally controls timely image processing according to the detection state of the object in the vehicle periphery monitoring device, and maintains its detection accuracy. However, it aims at reducing the calculation load of image processing.

  The present invention provides a vehicle periphery monitoring device. The periphery monitoring device is mounted on a vehicle and calculates an imaging unit that acquires an image around the vehicle, an object specifying unit that specifies an object around the vehicle from the image acquired by the imaging unit, and a movement vector of the object Movement vector calculation means for performing classification, determination means for determining the type of the object, and notification means for notifying the driver when the object is a predetermined target and the vehicle and the object are in a predetermined positional relationship, Prepare. The type determination means (i) performs the first type determination with high calculation load and high determination accuracy when the object exists in the first region where the arrival time until reaching the vehicle is equal to or shorter than a predetermined time. ii) When the object is present in the second area whose arrival time is longer than the predetermined time, the second type determination is performed with a lower calculation load and lower determination accuracy than the first type determination.

  According to the present invention, the accuracy of detecting a predetermined object around the vehicle is maintained high by increasing the accuracy of determining the type of an object that is relatively close to the vehicle and simultaneously reducing the calculation load for determining the type of an object that is relatively far away. However, the calculation load of the entire image processing can be reduced.

  According to an aspect of the present invention, the type determination unit is configured to perform a second operation on an object existing in the second area when an object exists in the first area and the second area, and a predetermined number of objects exist in the first area. The type determination is not performed, or the third type determination is performed with lower calculation load and lower determination accuracy than the second type determination.

  According to one aspect of the present invention, it is possible to concentrate calculation load on the determination of types of a plurality of objects that are relatively close to increase the detection accuracy, and to reduce the calculation load on the entire image processing.

  According to one aspect of the present invention, when it is determined from the arrival time for an object and the movement vector that the possibility that the object reaches the vehicle is greater than or equal to a predetermined value, the type determination unit does not determine the type for the object. .

  According to an aspect of the present invention, when an object with a high possibility of approaching a vehicle is detected, it is possible to promptly notify the driver by a warning or the like without performing the type determination.

It is a block diagram which shows the structure of the vehicle periphery monitoring apparatus according to one Example of this invention. It is a figure for demonstrating the attachment position of the infrared camera according to one Example of this invention. It is a figure which shows the processing flow in an image processing unit according to one Example of this invention. It is a figure for demonstrating the detection area ahead of the vehicle according to one Example of this invention. It is a figure which shows the processing flow of classification determination according to one Example of this invention. It is a figure which shows the processing flow of another classification determination according to one Example of this invention. It is a figure (table | surface) which shows the relationship between a detection area | region and alerting | reporting level according to one Example of this invention. It is a figure which shows the processing flow in another image processing unit according to one Example of this invention. It is a figure which shows the processing flow in another image processing unit according to one Example of this invention.

  Embodiments of the present invention will be described with reference to the drawings. FIG. 1 is a block diagram showing a configuration of a vehicle periphery monitoring device according to an embodiment of the present invention. The periphery monitoring device is mounted on a vehicle, and includes infrared cameras 1R and 1L, an image processing unit 12 for detecting an object around the vehicle based on image data captured by the infrared cameras 1R and 1L, and the detection result. A speaker 14 that generates an alarm based on sound or voice, a display device 16 that displays an image obtained through imaging by an infrared camera, and that causes a driver to recognize an object around the vehicle; Is provided. In a vehicle provided with a navigation device, corresponding functions provided in the navigation device may be used as the speaker 14 and the display device 16. A camera (such as a CCD camera) that uses another wavelength band (such as visible light) may be used instead of the infrared camera.

  The image processing unit 12 in FIG. 1 has the functions indicated by the blocks 121 to 124 as its configuration (function). That is, the image processing unit 12 includes an object specifying unit 121 that specifies an object around the vehicle from images acquired by the infrared cameras 1R and 1L, a movement vector calculating unit 122 that calculates a movement vector of the object, and an object type. Functions as a type determination unit 123 for determining the notification level, and a unit 124 for determining a notification level by the speaker 14 and the display device 16. The notification level determination unit 124 controls the speaker 14 and the display device 16 so as to notify the driver when the object is a predetermined target and the vehicle and the object are in a predetermined positional relationship.

  The function of each block is realized by a computer (CPU) included in the image processing unit 12. The configuration of the image processing unit 12 may be incorporated in the navigation device.

  The image processing unit 12 includes, for example, an A / D conversion circuit that converts an input analog signal into a digital signal, an image memory that stores a digitized image signal, and a central processing unit (CPU) that performs various arithmetic processes. ), RAM used by the CPU to store data for calculation, ROM to store programs executed by the CPU and data to be used (including tables and maps), drive signals for the speakers 14, display signals to the display device 16, etc. An output circuit for outputting is provided. The output signals of the infrared cameras 1R and 1L are converted into digital signals and input to the CPU.

  FIG. 2 is a view for explaining the mounting position of the infrared camera shown in FIG. 1 according to one embodiment of the present invention. As shown in FIG. 2, the infrared cameras 1R and 1L are disposed at the front portion of the vehicle 10 at positions that are substantially symmetrical with respect to the central axis in the lateral direction of the vehicle 10, and the optical axes of the two cameras 1R and 1L are mutually connected. They are parallel and fixed so that their height from the road surface is equal. The infrared cameras 1R and 1L have a characteristic that the output signal level increases (the luminance increases) as the temperature of the object increases. Reference numeral 2 in FIG. 2 represents a display screen when a head-up display (hereinafter referred to as “HUD”) is used as the display device 16. As shown in the figure, the HUD is provided such that the screen 2 is displayed at a front position of the driver on the front window of the vehicle 10.

  FIG. 3 is a process flow executed by the image processing unit 12 according to one embodiment of the present invention. This processing flow is executed at predetermined time intervals by the CPU of the image processing unit 12 calling a processing program stored in the memory. In the following description, a case where a black and white image is obtained by binarizing the acquired gray scale image is described as an example. However, multi-value quantization of three or more values may be performed. In that case, the number of threshold values to be set increases, but a multi-valued image can be obtained by performing basically the same processing as in the case of binarization.

  In step S10, an analog signal of an infrared image output from the infrared cameras 2R and 2L is input, and a grayscale image obtained by digitizing the analog signal by A / D conversion is stored in a memory. In step S10, a grayscale image (hereinafter referred to as a right image) obtained by the infrared camera 2R and a grayscale image (hereinafter referred to as a left image) obtained by the infrared camera 2L are acquired.

  In step S11, the right image is used as a reference image to perform binarization processing (processing in which a pixel having a luminance equal to or higher than a threshold is set to “1 (white) and a pixel smaller than the threshold is set to“ 0 (black) ”). A binary image is generated. In step S12, first, the image portion of each white area included in the binary image is converted into run length data (data of white pixel lines continuous in the x (horizontal) direction of the binary image). Next, a line having a portion overlapping in the y (vertical) direction of the binary image is labeled as one image portion, and the labeled image portion is extracted as an object image candidate.

  In the next step S13, the distance between the vehicle and the object candidate is calculated. Specifically, for example, the calculation is performed as follows.

  First, the gravity center G, the area S, and the aspect ratio ASPECT of the circumscribed rectangle of the image candidates of each object are calculated. Note that a specific calculation method is described in detail in, for example, Japanese Patent Application Laid-Open No. 2004-303219, and thus the description thereof is omitted here. Next, the image of the object extracted from the binary image based on the image captured by the infrared cameras 2R and 2L is determined at every predetermined sampling period, and the image determined to be the image of the same object Is stored in memory (time tracking). At the same time, the turning speed of the vehicle 10 is calculated by reading the vehicle speed detected by the vehicle speed sensor and the yaw rate detected by the yaw rate sensor and integrating the yaw rate over time.

  Further, one of the object image candidates tracked by the binary image of the reference image (right image) is selected, and the search image R1 (the circumscribed rectangle of the selected candidate image) is selected from the grayscale image of the right image. Image of the entire enclosed area) is extracted. Then, a search area for searching for an image corresponding to the search image R1 (hereinafter referred to as the corresponding image R1 ′) is set from the grayscale image of the left image, and a correlation calculation with the search image R1 is executed to determine the corresponding image R1 ′. Extract. Then, the difference between the centroid position of the search image R1 and the centroid position of the corresponding image R1 ′ is calculated as a parallax amount (number of pixels), and the distance z between the vehicle 10 and the object candidate is calculated based on the obtained parallax amount. . At the same time, the coordinates (x, y) and the distance z of the search image are converted into real space coordinates (X, Y, Z), and the coordinates of the real space position corresponding to the search image are calculated. The real space coordinates (X, Y, Z) are the origin 0 at the midpoint of the attachment position of the infrared cameras 2R, 2L in FIG. 2, X is the vehicle width direction of the vehicle 10, Y is the vertical direction, Z Is set in the forward direction of the vehicle 10.

  In the next step S14, the position of the object candidate is determined. FIG. 4 is a diagram for explaining a detection area in front of the vehicle according to one embodiment of the present invention. A triangular area 20 in FIG. 4 indicates an area that can be monitored by the infrared cameras 2 </ b> R and 2 </ b> L of the vehicle 10. The region 20 is divided into four regions A to D.

  Regions A and B are road regions, corresponding to a range obtained by adding a margin β (for example, about 50 to 100 cm) on both sides of the vehicle width α of the vehicle 10, in other words, the central axis in the lateral direction of the vehicle 10. This is a region having a width of (α / 2 + β) on both sides. These areas A and B are areas where the possibility of a collision is high if the object candidates continue to exist.

  Regions A and B are divided based on whether the time TTC (seconds) until the object candidate reaches the vehicle 10 is equal to or longer than the predetermined time T1 (seconds). Regions C and D are both roadside regions. These areas C and D are areas where the absolute value of the X coordinate is larger than those of the areas A and B (outside in the lateral direction of the areas A and B), and object candidates in these areas are directed to the vehicle 10 described later. This is an area where all intrusion collisions are possible.

  The object candidate distance and real space coordinates (X, Y, Z) obtained in step S13 are applied to the monitorable area in FIG. 4, and the object candidate is located in any of the areas A to D in the area 20. Determine whether. At this time, for example, it is first determined whether the area is on the road area (A + B) or the road area (C + D). Next, when it is in the road area (A + B), it is determined whether it is in the area A or B. A flag is set according to the determination result. For example, as illustrated in FIG. 4, Area_Flag is set to 1, 2, 4, and 4 depending on whether the area is located in any of the areas A, BC, and D, respectively.

  In the next step S15, the movement vector of the object candidate is calculated. Specifically, the turning angle correction is performed to correct the positional deviation on the image due to the turning of the vehicle 10 first. Next, the relative movement vector between the object candidate and the vehicle 10 is obtained from the time series data of the real space position of the same object candidate after the turning angle correction obtained from a plurality of images captured within a predetermined monitoring period. Is calculated. Note that a specific method for calculating the real space coordinates (X, Y, Z) and the movement vector is described in detail in, for example, Japanese Patent Application Laid-Open No. 2004-303219, and thus the description thereof is omitted here.

  When the calculated movement vector is starting from the areas C and D in FIG. 4 and heading toward the vehicle 10, there is a possibility of an intrusion collision toward the vehicle 10, so a flag is set. For example, as the vector indicated by reference numeral 21 in FIG. 4, 3 is set as Area_Flag for the movement vector of the target object candidate.

  In the next step S16, the object candidate type is determined. The type determination will be described with reference to FIGS. 5 and 6 are diagrams showing a processing flow for type determination according to one embodiment of the present invention.

  In step S20 of FIG. 5, it is determined whether or not the reachability of the object candidate to the vehicle 10 is greater than a predetermined value. Specifically, for example, the time TTC until each object candidate reaches the vehicle 10 is calculated from the position and movement vector of the object candidate obtained in steps S14 and S15 in FIG. 3, and the reciprocal (1 / TTC). Is a value AR indicating reachability. It is determined whether the AR is larger than a predetermined value. In other words, this determination corresponds to determining whether or not TTC is shorter than a predetermined time. Taking the monitorable area 20 in FIG. 4 as an example, the reachability of an object candidate located in an area where the TTC is smaller (close to the vehicle 10) in the area B where the TTC is 5 seconds or less exceeds a predetermined value. Is set as follows. Alternatively, when the size of the movement vector 21 in FIG. 4 is larger than a predetermined size, it is determined that the reachability of the object candidate exceeds a predetermined value.

  If the determination in step S20 is Yes, the type determination process is terminated and the process returns. This is because the possibility of collision between the object candidate and the vehicle 10 is high, so it is desirable to skip the type determination process and immediately proceed to the previous step to promptly notify the user. If this determination is No, the process proceeds to the next step S21.

  In step S21, it is determined whether or not an object candidate exists in the first region. Here, the first region means a region having a relatively high possibility of collision with the vehicle 10. Specifically, for example, it is determined whether or not an object candidate exists in the region B of FIG. 4, in other words, whether or not 2 is set as the Area_Flag of the object candidate. If this determination is Yes, the process proceeds to the next step S22, and if the determination is No, the process proceeds to step S23.

  In step S22, the first type determination is performed on the object candidates in the first area. The first type determination is a type determination with high calculation load and high determination accuracy. For example, it is determined whether an object candidate corresponds to a specific object such as a pedestrian. As a specific determination method, for example, when a pedestrian is targeted, luminance distribution, luminance profile, periodicity, or shape of a high-luminance portion of the object is a characteristic of the pedestrian ( Whether it is a pedestrian or not according to whether or not it falls under the head, shoulders, legs, etc.

  In step S23, it is determined whether or not the object candidate exists in the second region. Here, the second area means an area where the possibility of collision with the vehicle 10 is lower than that in the first area. Specifically, for example, it is determined whether or not an object candidate exists in the area A of FIG. 4, in other words, whether or not 1 is set as the Area_Flag of the object candidate. If this determination is Yes, the process proceeds to the next step S24, and if the determination is No, the process ends and returns.

  In step S24, the second type determination is performed on the object candidates in the second region. The second type determination is a type determination with a lower calculation load and lower determination accuracy than the first type determination. In other words, the second type determination is a simpler determination than the first type determination. Specifically, for example, the type of object candidate is determined by pattern matching with a small predetermined number of registered patterns in order to reduce the amount of calculation.

  Next, the processing flow of type determination in FIG. 6 will be described. The processing flow in FIG. 6 corresponds to a modified version of the processing flow in FIG. 5, and steps common to both flows are given the same step number. In FIG. 6, two steps S25 and S26 are added compared to FIG. Steps S20 to S24 are the same as those in FIG. 5, and the description thereof is omitted here.

  In step S25 of FIG. 6, it is determined whether or not there are a predetermined number or more object candidates in the first region. The meaning of the first area is the same as in step S21 in FIG. 5, and specifically, for example, whether or not there are a predetermined number or more of object candidates in area B in FIG. 4, in other words, 2 is set as Area_Flag. It is determined whether or not there are a predetermined number or more of object candidates. The predetermined number is set based on the arithmetic processing capability of the CPU of the image processing unit 12. If this determination is No, the process proceeds to step S24, and the second type determination is performed on the object candidates in the second region described in step S24 of FIG.

  If the determination in step S25 is Yes, in the next step S26, the third type determination is performed on the object candidate in the second region, or the determination is omitted and the process ends. Here, the third type determination is a type determination with lower calculation load and lower determination accuracy than the second type determination, and pattern matching with a smaller predetermined number (for example, 1, 2) of registered patterns (for example, only pedestrians). To determine the type of object candidate. By adopting this determination, when there are a plurality of objects in the first region where there is a high risk of collision, it is possible to concentrate the calculation load on the type processing and advance the image processing quickly. As a result, prompt notification to the driver is possible, and the risk of collision between the vehicle and the object can be further reduced.

  Returning to FIG. 3, in step S <b> 17, the reliability of the detected object, that is, the type determined object is determined. Specifically, the reliability of the object whose type is determined based on the number of times that the past type determination was correct, in other words, the number of times it was determined to be an object to be detected (for example, a pedestrian) Determine gender.

  In the next step S18, a notification level to the driver is determined, and a control signal is sent to the speaker 14 or the display device 16 serving as notification means. FIG. 7 is a diagram (table) showing the relationship between the detection area and the notification level according to one embodiment of the present invention. Info_Flag in FIG. 7 is a flag indicating the content of the notification, and Area_Flag is a flag indicating the position set in FIG.

  When the Area_Flag is 2 or 3, the risk of collision is high, so 2 is set as Info_Flag, and the speaker 14 determines whether to notify the driver of a sound (alarm) for alerting the driver. When Area_Flag is 1, 1 is set as Info_Flag, and a notification determination is made on the content of adding a frame to a predetermined object image displayed on the display device 16. When Area_Flag is 4, notification determination is made on the content of adding a frame to the object image only when the object is facing the vehicle. Thus, even if the position of the object is the same as the Z direction in FIG. 4, the content of the notification (warning) is switched depending on the position difference in the X direction. In addition, instead of adding a frame, or in addition to the frame, the content to be subjected to other highlight display (high brightness, blinking, color display, etc.) may be determined.

  Next, another example of the processing flow by the image processing unit according to one embodiment of the present invention will be described with reference to FIGS.

  The processing flow of FIG. 8 differs from the processing flow of FIG. 3 in that a determination step S150 is added between steps S15 and S16, and the other steps are the same as in FIG. In step S150 of FIG. 8, whether the detection of the object candidate for which the position is determined or the movement vector is obtained has already been confirmed in the past, that is, the detection cycle (processing cycle) before the predetermined number of times, in other words, the object is already detected. It is determined whether or not the type determination for is confirmed. If this determination is Yes, steps S16 to S18 are skipped and the process is terminated (reference numeral 22). By this skip processing, it is possible to avoid performing the determination processing again for a predetermined object that has already been subjected to the type determination, thereby reducing the calculation load on the CPU.

  In the processing flow of FIG. 9, compared to the processing flow of FIG. 3, step S14 is divided into two steps S140 and 144, and a determination step S142 is added between the two steps. Further, as in the case of FIG. The difference is that a determination step S150 is added between S15 and S16, and the other steps are the same as in FIG.

  In step S140 of FIG. 9, the object candidate position determination 1 is performed. Here, the position determination 1 determines whether the object candidate exists in the area A, C, or D in FIG. 4, in other words, whether Area_Flag is set to 1 or 4. That is, first, it is determined whether or not the object candidate exists in an area where the risk of collision with the vehicle 10 is low.

  In the next step S142, it is determined whether or not a sound notification (warning) has been made within a predetermined past time (predetermined detection cycle) for the object candidate whose position has been determined. If this determination is Yes, all subsequent steps are skipped and the process is terminated (reference numeral 23). This is because it is considered that the driver is paying attention to the object because the sound has already been notified within a predetermined time, and to reduce the annoyance caused by the continuous alarm. is there. As a result, it is possible to optimize the warning to the driver and reduce the calculation load of the CPU.

  If the determination in step S142 is No, the object candidate position determination 2 is performed in the next step S144. Here, the position determination 2 is performed when the object candidate exists in the area A in FIG. 4, that is, when Area_Flag 1 is set, whether the object candidate exists in the area B, in other words, Area_Flag is set to 2. Determine whether or not That is, it is determined whether or not the object candidate exists in an area where the risk of collision with the vehicle 10 is high.

  The subsequent step S15 and subsequent steps are the same as in FIG. 8. In step S150, it is determined whether or not the object candidate has been confirmed. If the determination is Yes, steps S16 to S18 are skipped and the process is terminated ( Reference numeral 23).

  The embodiment of the present invention has been described above, but the present invention is not limited to such an embodiment, and can be modified and used without departing from the spirit of the present invention.

  For example, in the above-described embodiment, the monitorable area in FIG. 4 is divided into four areas A to D, but the division method is not limited to this, and a division method such as division into three in the Z direction is adopted. May be. Further, the calculation load may be reduced (accuracy is reduced) as the distance in the X direction is longer (regions C and D in FIG. 4).

1R, 1L Infrared camera 2 Display screen (HUD)
DESCRIPTION OF SYMBOLS 10 Vehicle 12 Image processing unit 14 Speaker 16 Display apparatus 20 Area | region which can be monitored with an infrared camera

Claims (3)

An imaging means mounted on a vehicle for acquiring an image around the vehicle;
Object specifying means for specifying objects around the vehicle from the image acquired by the imaging means;
Movement vector calculation means for calculating a movement vector of the object;
Type determination means for determining the type of the object;
A vehicle periphery monitoring device comprising: a notification unit configured to notify a driver when the object is a predetermined target and the vehicle and the object are in a predetermined positional relationship,
The type determination means performs (i) a first type determination with high calculation load and high determination accuracy when the object exists in a first region where the arrival time until reaching the vehicle is a predetermined time or less. (Ii) When the object is present in the second region where the arrival time is longer than a predetermined time, the second type determination is performed with a lower calculation load and lower determination accuracy than the first type determination. Monitoring device.   The type determination means performs the second type determination for an object existing in the second area when an object exists in the first area and the second area and a predetermined number or more objects exist in the first area. The periphery monitoring device according to claim 1, wherein the third type determination is not performed or the third type determination is performed with lower calculation load and lower determination accuracy than the second type determination.   The type determination unit does not determine the type of the object when it is determined from the arrival time of the object and the movement vector that the possibility that the object reaches the vehicle is a predetermined value or more. The periphery monitoring apparatus according to 1 or 2.

Images