Pedestrian collision determination system, pedestrian collision determination device, pedestrian collision determination method, and program

Abstract

Provided is a pedestrian collision determination system, comprising: an image capture device (102) that is mounted in a vehicle (101) and captures images of the periphery of the vehicle; a first detection unit (201) that detects the direction of advance of the vehicle; a second detection unit (202) that detects the ground-contact position of a pedestrian captured by the image capture device; a third detection unit (203) that detects the orientation of the pedestrian captured by the image capture device; a possible collision area setting unit (204) that sets a possible collision area having a shape and a size corresponding to the environment when the pedestrian was captured by the image capture device; and a determination unit (205) that determines the possibility of a collision between the vehicle and the pedestrian captured in the image, on the basis of the direction of advance of the vehicle, the ground-contact position of the pedestrian, the orientation of the pedestrian, and the possible collision area.

Description

Pedestrian collision detection system, pedestrian collision detection device, pedestrian collision detection method, and computer-readable medium

[0001] The present disclosure relates to a pedestrian collision determination system, a pedestrian collision determination device, a pedestrian collision determination method, and a computer-readable medium.

[0002] Patent Document 1 describes an invention that predicts at an early stage whether a pedestrian on the side of the road in the lane in which the vehicle is traveling may cross the road, and assists in preventing a collision with the pedestrian. The invention described in Patent Document 1 determines whether a pedestrian on the opposite side of the road in the oncoming lane is crossing the road, based on whether the pedestrian in front recognizes that the oncoming pedestrian is crossing from the direction of the pedestrian's face, and whether the pedestrian recognizes the presence of the vehicle.

[0003] Patent Document 2 describes an invention that appropriately supports a driver's driving when there is a possibility of a collision between the vehicle and a pedestrian. The invention described in Patent Document 2 captures an image of the pedestrian and recognizes the direction of the pedestrian's face. The invention described in Patent Document 2 also estimates whether there is a possibility of a collision between the vehicle and the pedestrian based on the pedestrian's movement. The invention described in Patent Document 2 also determines whether the presence of the vehicle is recognized based on the direction of the face when it is estimated that there is a possibility of a collision. The invention described in Patent Document 2 then controls an alarm device depending on the determination result.

[0004] Patent Document 3 describes an invention that improves the accuracy of identifying a subject that appears in an image acquired by an imaging device. In the invention described in Patent Document 3, a processing unit performs processing on image data in which a skeletal model in which a central feature point corresponding to the center of the human body is connected to a plurality of feature points corresponding to the limbs is applied to the subject.

[0005] JP 2012-234499 A JP 2014-059841 A JP 2021-060814 A

[0006] However, Patent Documents 1 to 3 do not mention in detail a method for setting a potential collision area where a vehicle and a person may collide. Therefore, an object of the present disclosure is to provide a pedestrian collision determination system that changes the shape and size of the potential collision area depending on the environment.

[0007] A pedestrian collision determination system according to the present disclosure includes: an imaging device mounted on a vehicle and capturing images of the surroundings of the vehicle; a first detection means for detecting the traveling direction of the vehicle; a second detection means for detecting the ground contact position of the pedestrian captured by the imaging device; a third detection means for detecting the orientation of the pedestrian captured by the imaging device; a possible collision area setting means for setting a possible collision area having a shape and size according to the environment when the pedestrian is captured by the imaging device; and a determination means for determining the possibility of a collision between the vehicle and the captured pedestrian based on the traveling direction of the vehicle detected by the first detection means, the ground contact position of the captured pedestrian detected by the second detection means, the orientation of the captured pedestrian detected by the third detection means, and the possible collision area set by the possible collision area setting means.

[0008] A pedestrian collision determination device according to the present disclosure includes: an imaging device mounted on a vehicle and capturing images of the surroundings of the vehicle; a first detection means for detecting the traveling direction of the vehicle; a second detection means for detecting the ground contact position of the pedestrian captured by the imaging device; a third detection means for detecting the orientation of the pedestrian captured by the imaging device; a possible collision area setting means for setting a possible collision area having a shape and size according to the environment when the pedestrian is captured by the imaging device; and a determination means for determining the possibility of a collision between the vehicle and the captured pedestrian based on the traveling direction of the vehicle detected by the first detection means, the ground contact position of the captured pedestrian detected by the second detection means, the orientation of the captured pedestrian detected by the third detection means, and the possible collision area set by the possible collision area setting means.

[0009] The pedestrian collision determination method of the present disclosure is a pedestrian collision determination method that captures an image of the area around a vehicle, detects the traveling direction of the vehicle, detects the ground contact position of the imaged pedestrian, detects the orientation of the imaged pedestrian, sets a potential collision area having a shape and size according to the environment when the pedestrian was imaged, and determines the possibility of a collision between the vehicle and the imaged pedestrian based on the detected traveling direction of the vehicle, the detected ground contact position of the imaged pedestrian, the detected orientation of the imaged pedestrian, and the set potential collision area.

[0010] The computer-readable medium of the present disclosure is a non-transitory computer-readable medium storing a program that causes an information processing device to perform the following steps: capture an image of the surroundings of a vehicle; detect the direction of travel of the vehicle; detect the ground contact position of the imaged pedestrian; detect the orientation of the imaged pedestrian; set a potential collision area having a shape and size according to the environment when the pedestrian was imaged; and determine the possibility of a collision between the vehicle and the imaged pedestrian based on the detected direction of travel of the vehicle, the detected ground contact position of the imaged pedestrian, the detected orientation of the imaged pedestrian, and the set potential collision area.

[0011] The present disclosure provides a pedestrian collision determination system that changes the shape and size of a potential collision area depending on the environment.

[0012] Fig. 1 is a schematic diagram of a pedestrian collision determination system according to an embodiment; Fig. 2 is a block diagram showing the configuration of a pedestrian collision determination system according to an embodiment; Fig. 3 is a flowchart of a pedestrian collision determination method according to an embodiment; Fig. 4 is a diagram showing a method for detecting a pedestrian's direction according to an embodiment; Fig. 5 is a diagram showing an example of a pedestrian's ground contact position and a pedestrian collision possibility area according to an embodiment;

[0013] Embodiments Hereinafter, embodiments of the present invention will be described with reference to the drawings. However, the invention according to the claims is not limited to the following embodiments. Furthermore, not all of the configurations described in the embodiments are necessarily essential as means for solving the problems. For clarity of explanation, the following description and drawings have been omitted and simplified as appropriate. In each drawing, the same elements are given the same reference numerals, and repeated explanations are omitted as necessary.

[0014] (Description of a pedestrian collision determination system according to an embodiment) Fig. 1 is a schematic diagram of a pedestrian collision determination system according to an embodiment. Fig. 2 is a block diagram showing the configuration of a pedestrian collision determination system according to an embodiment. The pedestrian collision determination system according to an embodiment will be described with reference to Figs. 1 and 2 .

[0015] 1, the pedestrian collision detection system 100 is a system used for automatic driving or driving assistance of a vehicle 101. The pedestrian collision detection system 100 is used to brake the vehicle or issue a warning to the driver or a remote monitor when it determines that there is a high possibility that the vehicle will collide with a pedestrian. The pedestrian collision detection system 100 includes at least an imaging device 102 and an information processing device 103 mounted on the vehicle 101.

[0016] The vehicle 101 is a moving body that carries people, such as a bus or truck. The vehicle 101 may be driven autonomously or by a driver. The vehicle 101 may also be an unmanned moving body that does not carry people.

[0017] The imaging device 102 is an exterior camera mounted on a vehicle that captures the environment around the vehicle. The imaging device 102 captures images of pedestrians or roads around the vehicle. The imaging device 102 is connected to a traffic management / monitoring center via a wireless network. The wireless network may be, for example, a mobile communication network such as 4G (4th Generation), LTE (Long Term Evolution), or 5G (5th Generation), or Wi-Fi (registered trademark). The imaging device 102 distributes the captured images to the traffic management / monitoring center.

[0018] Furthermore, the imaging device 102 may use one imaging device, two or three imaging devices, or four or more imaging devices depending on the angle of view. By using a wide-angle imaging device in front, the surrounding environment can be captured with one imaging device. The surrounding environment can be captured by combining images captured by two imaging devices facing in two directions, the front right and the front left. Furthermore, the surrounding environment can be captured by combining images captured by multiple imaging devices facing in three directions, the right, center, and left. Similarly, the surrounding environment can be captured by combining images captured by multiple imaging devices facing in four or more directions.

[0019] The vehicle 101 may be equipped with an in-vehicle camera 104. The in-vehicle camera 104 is an imaging device used to capture images of the interior of the vehicle and prevent or respond to trouble inside the vehicle. Trouble inside the vehicle may include a passenger staying behind or a passenger falling over. The in-vehicle camera 104 is also connected to the operation management / monitoring center via a wireless network. The in-vehicle camera 104 distributes the captured images to the operation management / monitoring center.

[0020] The traffic management / monitoring center is a location located outside the vehicle that manages and monitors the operation of the vehicle. The traffic management / monitoring center collects information from multiple vehicles and simultaneously manages and monitors the multiple vehicles. The traffic management / monitoring center instructs personnel to rush to the scene in an emergency based on images captured by the imaging device 102 or the in-vehicle camera 104. The traffic management / monitoring center also includes an information processing device 103. The information processing device 103 analyzes the situation inside and outside the vehicle using artificial intelligence (AI).

[0021] The information processing device 103 includes a memory for storing a program and a processor for executing the program. The information processing device 103 may be configured as a single information processing device or may be configured as multiple information processing devices. The information processing device 103 may have some or all of its functions executed by the cloud. Here, the information processing device 103 is installed in an operation management / monitoring center, but some or all of the information processing device 103 may be installed in the vehicle 101. When all of the information processing devices 103 are installed in the vehicle 101, the vehicle 101 can be said to function as one pedestrian collision detection device.

[0022] 2 , the pedestrian collision determination system 100 includes a vehicle 101, an imaging device 102, a first detection unit 201, a second detection unit 202, a third detection unit 203, a collision possibility area setting unit 204, and a determination unit 205. Here, the information processing device 103 executes the functions of the first detection unit 201, the second detection unit 202, the third detection unit 203, the collision possibility area setting unit 204, and the determination unit 205.

[0023] The first detection unit 201 detects the traveling direction of the vehicle 101. The first detection unit 201 detects the traveling direction of the vehicle from the driver's steering operation and blinker operation.

[0024] The second detection unit 202 detects the ground contact position of the pedestrian captured by the imaging device 102. The second detection unit 202 pays particular attention to the feet of the pedestrian around the vehicle 101, and detects the ground contact position of the pedestrian on the road.

[0025] The third detection unit 203 detects the direction of the pedestrian captured by the imaging device 102. The third detection unit 203 detects the direction in which the pedestrian around the vehicle 101 is facing.

[0026] The potential collision area setting unit 204 sets a potential collision area having a shape and size according to the environment when the pedestrian is imaged. The potential collision area having a size and shape according to the environment is an area where a vehicle may collide, and the shape and size of the front, rear, left, and right sides change according to the surrounding road environment. The environment includes the vehicle speed, the vehicle acceleration, road signs around the vehicle, and road markings around the vehicle.

[0027] The determination unit 205 determines the possibility of a collision between the vehicle 101 and the pedestrian based on the traveling direction of the vehicle 101, the pedestrian's touchdown position, the pedestrian's orientation, and the possible collision area set by the possible collision area setting unit. The determination unit 205 determines that there is a possibility of a collision when the traveling direction of the vehicle 101 from the position of the vehicle 101 and the orientation of the pedestrian from the pedestrian's touchdown position intersect. Furthermore, the determination unit 205 determines that there is a high possibility of a collision when the pedestrian's touchdown position is within the possible collision area.

[0028] The first detection unit 201 may also be referred to as first detection means. The second detection unit 202 may also be referred to as second detection means. The third detection unit 203 may also be referred to as third detection means. The potential collision area setting unit 204 may also be referred to as potential collision area setting means. The determination unit 205 may also be referred to as determination means.

[0029] In this way, a pedestrian collision determination system can be provided that changes the shape and size of the potential collision area depending on the environment.

[0030] (Description of Pedestrian Collision Determination Method According to the Embodiment) Fig. 3 is a flowchart of a pedestrian collision determination method according to the embodiment. The pedestrian collision determination method according to the embodiment will be described with reference to Fig. 3 .

[0031] First, the surroundings of the vehicle 101 are imaged (step S301). The imaging device 102 images the environment around the vehicle 101 and pedestrians. The information processing device 103 acquires images of the pedestrians. Next, the traveling direction of the vehicle 101 is detected (step S302). The first detection unit 201 detects the traveling direction of the vehicle 101. Next, the ground contact position of the pedestrian is detected (step S303). The second detection unit 202 detects the ground contact position of the imaged pedestrian. Next, the direction of the pedestrian is detected (step S304). The third detection unit 203 detects the direction of the imaged pedestrian.

[0032] Next, a potential collision area is set (step S305). The potential collision area setting unit 204 sets a potential collision area having a shape and size according to the environment when the pedestrian is imaged. Next, the various conditions are combined to determine the possibility of a collision between the vehicle 101 and the pedestrian (step S306), and the process ends. The determination unit 205 determines the possibility of a collision between the vehicle 101 and the pedestrian based on the traveling direction of the vehicle 101, the pedestrian's ground contact position, the pedestrian's orientation, and the potential collision area. If there is a possibility of a collision between the vehicle 101 and the pedestrian, an instruction to assist the driving of the vehicle 101, such as preparing the vehicle 101 to apply the brakes, can be issued. Alternatively, the vehicle 101 may be made to apply the brakes.

[0033] In this way, a pedestrian detection method can be provided that changes the shape and size of the potential collision area depending on the environment.

[0034] (Description of a method for detecting the direction of a pedestrian according to an embodiment) Fig. 4 is a diagram showing a method for detecting the direction of a pedestrian according to an embodiment. The method for detecting the direction of a pedestrian according to an embodiment will be described with reference to Fig. 4 .

[0035] As shown in the left image of Figure 4, skeletal feature points of a pedestrian are extracted from an image of the pedestrian. The orientation of the pedestrian is detected based on the state of the pedestrian's skeleton. For example, feature points are extracted such as the right eye, left eye, right ear, left ear, nose, neck, right shoulder, left shoulder, right elbow, left elbow, right wrist, left wrist, right hip, left hip, right knee, left knee, right ankle, and left ankle. Then, as shown in the second image from the left in Figure 4, for a person facing the vehicle, the eye and nose feature points are recognized. Also, for a person facing the vehicle, the right skeletal feature points appear on the left, and the left skeletal feature points appear on the right. As shown in the middle image of Figure 4, for a person facing away from the vehicle, the eye and nose feature points are not recognized. Also, for a person facing away from the vehicle, the right skeletal feature points appear on the right, and the left skeletal feature points appear on the left.

[0036] As shown in the second image from the right in Figure 4, the eyes and nose of a person facing left as viewed from the vehicle are recognized to the left of their neck. Also, for a person facing left as viewed from the vehicle, the feature points of the right side of the skeleton appear on the left, and the feature points of the left side of the skeleton appear on the right. As shown in the right image in Figure 4, the eyes and nose of a person facing right as viewed from the vehicle are recognized to the right of their neck. Also, for a person facing right as viewed from the vehicle, the feature points of the right side of the skeleton appear on the right, and the feature points of the left side of the skeleton appear on the left.

[0037] In this way, the direction of a pedestrian can be detected by using AI that inputs the characteristic points of the pedestrian's skeleton and outputs the direction in which the pedestrian is facing.

[0038] (Explanation of Setting of Possible Collision Area According to the Embodiment) Fig. 5 is a diagram showing an example of a pedestrian's landing position and a pedestrian collision possibility area according to the embodiment. Setting of the possible collision area will be described with reference to Fig. 5 .

[0039] The upper diagram of Figure 5 shows the surrounding environment captured by the left imaging device out of three imaging devices mounted on a vehicle, for example, on the right, left, and center. As shown in the upper diagram of Figure 5, if a person is facing right from the vehicle, the direction of travel of the vehicle and the pedestrian cross, which could result in a collision between the vehicle and the pedestrian. If the person is facing right and their ground contact point is in a potential collision area, it is determined that there is a high possibility of collision. In this case, the control of the vehicle may be affected. On the other hand, if the person is facing right and their ground contact point is not in a potential collision area, it is determined that there is a low possibility of collision.

[0040] The shape and size of the potential collision area vary depending on the surrounding environment. Specific examples are shown in the center and bottom diagrams of FIG. 5. The center diagram of FIG. 5, like the top diagram, is an image captured by an imaging device installed on the left side of the vehicle. As shown in the left diagram of the center diagram of FIG. 5, when the vehicle is traveling in the left lane and the vehicle traveling speed is slow, the potential collision area spreads out from the center of the vehicle and extends forward and to the left, but the area is not very large. On the other hand, as shown in the right diagram of the center diagram of FIG. 5, when the vehicle is traveling in the left lane and the vehicle traveling speed is fast, the potential collision area spreads out from the center and extends significantly forward. This is because the potential collision area extends forward when the vehicle is fast. It is preferable that the potential collision area be shaped and sized so that it is larger forward as the vehicle is faster and smaller forward as the vehicle is slower.

[0041] Furthermore, as shown in the right diagram in the lower diagram of Figure 5, when a vehicle is driving on the left lane of a road with left-hand traffic, the potential collision area spreads out from the center and extends forward and to the left, but the area is not very large. On the other hand, as shown in the right diagram in the lower diagram of Figure 5, when a vehicle is driving on the left lane and is driving in front of a crosswalk, the potential collision area spreads out from the center and extends significantly forward and to the left. This is because there is a high possibility of pedestrians entering from the left near a crosswalk. Here, the left side is used as an example, but it can also be the right side when driving on the right side, or more accurately, toward the side of the road. The same applies to intersections, railroad crossings, and traffic lights.

[0042] Although not shown, if there is a bus stop near the vehicle, the potential collision area preferably has a shape and size that includes an area extending from the vehicle to the bus stop, because there is a high possibility that people will be waiting near the bus stop.

[0043] Although not shown, when a vehicle turns right, it is preferable that the potential collision area has a larger shape and size on the right side because there is a higher possibility of a person entering from the right side. Similarly, when a vehicle turns left, it is preferable that the potential collision area has a larger shape and size on the left side because there is a higher possibility of a person entering from the left side.

[0044] Furthermore, although not shown, it is preferable that the potential collision area has a shape and size that correspond to the gradient of the road on which the vehicle is traveling. When the vehicle is traveling uphill, it is preferable that the potential collision area has a shape and size that is smaller in the forward direction in accordance with the range of visibility ahead. When the vehicle is traveling uphill, it is easier to stop the vehicle, and the potential collision area can be made smaller in the forward direction. It can also be said that the gradient of the uphill slope is proportional to the range of visibility ahead. Similarly, when the vehicle is traveling downhill, it is preferable that the potential collision area has a shape and size that is larger in the forward direction in accordance with the range of visibility ahead. When the vehicle is traveling downhill, it is harder to stop the vehicle, and the potential collision area must be made larger in the forward direction. It can also be said that the gradient of the downhill slope is proportional to the range of visibility ahead.

[0045] In this way, the pedestrian collision potential area can be shaped and sized according to the environment around the vehicle.

[0046] Furthermore, part or all of the processing in the information processing device 103 described above can be realized as a computer program. Such a program can be stored using various types of non-transitory computer-readable media and supplied to a computer. Non-transitory computer-readable media include various types of tangible recording media. Examples of non-transitory computer-readable media include magnetic recording media (e.g., flexible disks, magnetic tapes, hard disk drives), magneto-optical recording media (e.g., magneto-optical disks), CD-ROMs (Read Only Memory), CD-Rs, CD-R / Ws, and semiconductor memories (e.g., mask ROMs, PROMs (Programmable ROMs), EPROMs (Erasable PROMs), flash ROMs, and RAMs (Random Access Memory)). Furthermore, the program may be supplied to a computer by various types of temporary computer-readable media. Examples of temporary computer-readable media include electrical signals, optical signals, and electromagnetic waves. The temporary computer-readable medium can supply the program to the computer via a wired communication path such as an electric wire or an optical fiber, or via a wireless communication path.

[0047] The present invention is not limited to the above-described embodiment, and can be modified as appropriate within the scope of the invention.

[0048] Some or all of the above embodiments may be described as in the following supplementary notes, but are not limited to the following: (Supplementary Note 1) A pedestrian collision determination system comprising: an imaging device mounted on a vehicle and capturing images of the surroundings of the vehicle, a first detection means for detecting a traveling direction of the vehicle, a second detection means for detecting a ground contact position of a pedestrian captured by the imaging device, a third detection means for detecting a direction of the pedestrian captured by the imaging device, a possible collision area setting means for setting a possible collision area having a shape and size according to the environment when the pedestrian is captured by the imaging device, and a determination means for determining a possibility of a collision between the vehicle and the captured pedestrian based on the traveling direction of the vehicle detected by the first detection means, the ground contact position of the captured pedestrian detected by the second detection means, the direction of the captured pedestrian detected by the third detection means, and the possible collision area set by the possible collision area setting means. (Supplementary Note 2) The pedestrian collision determination system according to Supplementary Note 1, wherein the determination means determines that a collision is highly likely when a direction of travel of the vehicle from the position of the vehicle and a direction of the pedestrian from a contact point of the pedestrian intersect and when the contact point of the pedestrian is within the collision possibility area. (Supplementary Note 3) The pedestrian collision determination system according to Supplementary Note 1 or 2, wherein the collision possibility area setting means sets a collision possibility area that has a shape and size that is larger in front of, as seen from, the vehicle as the speed of the vehicle is faster and that is smaller in front of, as seen from the vehicle as the speed of the vehicle is slower. (Supplementary Note 4) The pedestrian collision determination system according to Supplementary Note 1 or 2, wherein the collision possibility area setting means sets a collision possibility area that has a large shape and size toward the roadside and in front of, as seen from the vehicle, when the vehicle is in front of a crosswalk, intersection, railroad crossing, or traffic light. (Supplementary Note 5) The pedestrian collision determination system according to Supplementary Note 1 or 2, wherein the collision possibility area setting means sets a collision possibility area having a shape and size including an area extending from the vehicle to a bus stop when there is a bus stop near the vehicle.(Supplementary Note 6) The pedestrian collision determination system according to Supplementary Note 1 or 2, wherein the potential collision area setting means sets a potential collision area having a shape and size that is larger on the right side when the vehicle turns right and larger on the left side when the vehicle turns left. (Supplementary Note 7) The pedestrian collision determination system according to Supplementary Note 1 or 2, wherein the potential collision area setting means sets a potential collision area having a shape and size according to the gradient of a road on which the vehicle is traveling. (Supplementary Note 8) The pedestrian collision determination system according to Supplementary Note 1, wherein the third detection means detects the orientation of the pedestrian based on the state of the pedestrian's skeleton. (Supplementary Note 9) A pedestrian collision determination device comprising: an imaging device mounted on a vehicle and capturing images of the surroundings of the vehicle, a first detection means for detecting a traveling direction of the vehicle, a second detection means for detecting a ground contact position of a pedestrian captured by the imaging device, a third detection means for detecting an orientation of the pedestrian captured by the imaging device, a possible collision area setting means for setting a possible collision area having a shape and size according to an environment when the pedestrian is captured by the imaging device, and a determination means for determining a possibility of a collision between the vehicle and the captured pedestrian based on the traveling direction of the vehicle detected by the first detection means, the ground contact position of the captured pedestrian detected by the second detection means, the orientation of the captured pedestrian detected by the third detection means, and the possible collision area set by the possible collision area setting means. (Supplementary Note 10) The pedestrian collision determination device according to Supplementary Note 9, wherein the determination means determines that a collision is highly likely when the traveling direction of the vehicle from a position of the vehicle and the orientation of the pedestrian from the ground contact position of the pedestrian intersect and when the ground contact position of the pedestrian is within the possible collision area. (Supplementary Note 11) The pedestrian collision determination device according to Supplementary Note 9 or 10, wherein the collision possibility area setting means sets a collision possibility area having a shape and size that is larger in front of the vehicle as seen from the vehicle when the vehicle speed is faster, and a shape and size that is smaller in front of the vehicle as seen from the vehicle when the vehicle speed is slower.(Supplementary Note 12) The pedestrian collision determination device according to Supplementary Note 9 or 10, wherein the possible collision area setting means sets a possible collision area having a large shape and size towards the roadside and ahead as seen from the vehicle when the vehicle is in front of a crosswalk, intersection, railroad crossing, or traffic light. (Supplementary Note 13) The pedestrian collision determination device according to Supplementary Note 9 or 10, wherein the possible collision area setting means sets a possible collision area having a shape and size including an area extending from the vehicle to a bus stop when there is a bus stop near the vehicle. (Supplementary Note 14) The pedestrian collision determination device according to Supplementary Note 9 or 10, wherein the possible collision area setting means sets a possible collision area having a large shape and size on the right side when the vehicle is turning right, and a large shape and size on the left side when the vehicle is turning left. (Supplementary Note 15) The pedestrian collision determination device according to Supplementary Note 9 or 10, wherein the possible collision area setting means sets a possible collision area having a shape and size according to the gradient of the road on which the vehicle is traveling. (Supplementary Note 16) The pedestrian collision determination device according to Supplementary Note 9, wherein the third detection means detects the orientation of the pedestrian based on the state of the pedestrian's skeleton. (Supplementary Note 17) A pedestrian collision determination method comprising: capturing an image of the surroundings of a vehicle; detecting a traveling direction of the vehicle; detecting a ground contact position of the captured pedestrian; detecting an orientation of the captured pedestrian; setting a potential collision area having a shape and size according to the environment when the pedestrian was imaged; and determining a possibility of a collision between the vehicle and the captured pedestrian based on the detected traveling direction of the vehicle, the detected ground contact position of the captured pedestrian, the detected orientation of the captured pedestrian, and the set potential collision area. (Supplementary Note 18) The pedestrian collision determination method according to Supplementary Note 17, wherein a collision is determined to be highly likely when the traveling direction of the vehicle from the position of the vehicle and the orientation of the pedestrian from the ground contact position of the pedestrian intersect and the pedestrian's ground contact position is within the potential collision area. (Supplementary Note 19) A pedestrian collision determination method according to Supplementary Note 17 or 18, in which a collision possibility area is set that has a shape and size that is larger in front of the vehicle as seen from the vehicle when the vehicle speed is faster, and a shape and size that is smaller in front of the vehicle as seen from the vehicle when the vehicle speed is slower.(Supplementary Note 20) The pedestrian collision determination method according to Supplementary Note 17 or 18, wherein a potential collision area having a large shape and size is set toward the roadside and ahead as seen from the vehicle when the vehicle is in front of a crosswalk, intersection, railroad crossing, or traffic light. (Supplementary Note 21) The pedestrian collision determination method according to Supplementary Note 17 or 18, wherein a potential collision area having a shape and size including an area extending from the vehicle to a bus stop when the vehicle is near the bus stop. (Supplementary Note 22) The pedestrian collision determination method according to Supplementary Note 17 or 18, wherein a potential collision area having a large shape and size is set to the right when the vehicle is turning right, and a large shape and size is set to the left when the vehicle is turning left. (Supplementary Note 23) The pedestrian collision determination method according to Supplementary Note 17 or 18, wherein a potential collision area having a shape and size according to the gradient of the road on which the vehicle is traveling is set. (Supplementary Note 24) The pedestrian collision determination method according to Supplementary Note 17, wherein the orientation of the pedestrian is detected based on the state of the pedestrian's skeleton. (Supplementary Note 25) A non-transitory computer-readable medium storing a program that causes an information processing device to perform the following steps: capture an image of the surroundings of a vehicle, detect a traveling direction of the vehicle, detect a touchdown position of the imaged pedestrian, detect an orientation of the imaged pedestrian, set a potential collision area having a shape and size according to the environment when the pedestrian was imaged, and determine a possibility of a collision between the vehicle and the imaged pedestrian based on the detected traveling direction of the vehicle, the detected touchdown position of the imaged pedestrian, the detected orientation of the imaged pedestrian, and the set potential collision area. (Supplementary Note 26) A non-transitory computer-readable medium storing the program according to Supplementary Note 25, which determines that a collision is highly likely when the traveling direction of the vehicle from the position of the vehicle and the orientation of the pedestrian from the touchdown position of the pedestrian intersect and the pedestrian's touchdown position is within the potential collision area. (Supplementary Note 27) A non-transitory computer-readable medium storing a program according to Supplementary Note 25 or 26, which sets a potential collision area to a shape and size that is larger in the forward direction depending on the speed when the speed of the vehicle is fast, and smaller in the forward direction depending on the speed when the speed of the vehicle is slow.(Supplementary Note 28) A non-transitory computer-readable medium storing the program of Supplementary Note 25 or 26, which sets a potential collision area having a large shape and size toward the roadside and ahead as seen from the vehicle when the vehicle is in front of a crosswalk, intersection, railroad crossing, or traffic light. (Supplementary Note 29) A non-transitory computer-readable medium storing the program of Supplementary Note 25 or 26, which sets a potential collision area having a shape and size that includes an area extending from the vehicle to a bus stop when there is a bus stop near the vehicle. (Supplementary Note 30) A non-transitory computer-readable medium storing the program of Supplementary Note 25 or 26, which sets a potential collision area having a shape and size that is large on the right side when the vehicle is turning right, and large on the left side when the vehicle is turning left. (Supplementary Note 31) A non-transitory computer-readable medium storing the program of Supplementary Note 25 or 26, which sets a potential collision area having a shape and size that corresponds to the gradient of the road on which the vehicle is traveling. (Supplementary Note 32) A non-transitory computer-readable medium storing the program according to Supplementary Note 25, which detects the orientation of the pedestrian based on a state of the pedestrian's skeleton.

[0049] REFERENCE SIGNS LIST 100 Pedestrian collision determination system, 101 Vehicle, 102 Imaging device, 103 Information processing device, 201 First detection unit, 202 Second detection unit, 203 Third detection unit, 204 Possible collision area setting unit, 205 Determination unit

Claims

1. An imaging device mounted on a vehicle for capturing an image of an area around the vehicle; A first detection means for detecting a traveling direction of the vehicle; a second detection means for detecting a ground contact position of a pedestrian captured by the imaging device; a third detection means for detecting a direction of a pedestrian captured by the imaging device; a collision possibility area setting means for setting a collision possibility area having a shape and size according to the environment when a pedestrian is imaged by the imaging device; a determination means for determining a possibility of a collision between the vehicle and the imaged pedestrian based on a traveling direction of the vehicle detected by the first detection means, a ground contact position of the imaged pedestrian detected by the second detection means, a direction of the imaged pedestrian detected by the third detection means, and the possible collision area set by the possible collision area setting means.

2. 2. The pedestrian collision determination system according to claim 1, wherein the determination means determines that the collision possibility is high when a traveling direction of the vehicle from the position of the vehicle intersects with a direction of the pedestrian from a landing position of the pedestrian and when the landing position of the pedestrian is within the collision possibility area.

3. 3. The pedestrian collision determination system according to claim 1, wherein the collision possibility area setting means sets a collision possibility area having a shape and size that is larger in front of the vehicle as seen from the vehicle when the vehicle speed is faster and a shape and size that is smaller in front of the vehicle as seen from the vehicle when the vehicle speed is slower.

4. 3. The pedestrian collision determination system according to claim 1, wherein the collision possibility area setting means sets a collision possibility area having a large shape and size in a roadside direction and ahead as seen from the vehicle when the vehicle is in front of a crosswalk, an intersection, a railroad crossing, or a traffic light.

5. 3. The pedestrian collision determination system according to claim 1, wherein the collision possibility area setting means sets a collision possibility area having a shape and size including an area extending from the vehicle to a bus stop when the bus stop is located near the vehicle.

6. 3. The pedestrian collision determination system according to claim 1, wherein the collision possibility area setting means sets a collision possibility area having a shape and size larger on the right side when the vehicle turns right and larger on the left side when the vehicle turns left.

7. 3. The pedestrian collision determination system according to claim 1, wherein the collision possibility area setting means sets a collision possibility area having a shape and size according to a gradient of a road on which the vehicle is traveling.

8. An imaging device mounted on a vehicle for capturing an image of an area around the vehicle; A first detection means for detecting a traveling direction of the vehicle; a second detection means for detecting a ground contact position of a pedestrian captured by the imaging device; a third detection means for detecting a direction of a pedestrian captured by the imaging device; a collision possibility area setting means for setting a collision possibility area having a shape and size according to the environment when a pedestrian is imaged by the imaging device; a determination means for determining a possibility of a collision between the vehicle and the imaged pedestrian based on a traveling direction of the vehicle detected by the first detection means, a ground contact position of the imaged pedestrian detected by the second detection means, a direction of the imaged pedestrian detected by the third detection means, and the possible collision area set by the possible collision area setting means.

9. Capture images of the vehicle's surroundings, Detecting a traveling direction of the vehicle; Detecting a landing position of the imaged pedestrian; Detecting the orientation of the captured pedestrian; setting a collision possibility area having a shape and size according to the environment when the pedestrian is imaged; A pedestrian collision determination method that determines the possibility of a collision between the vehicle and the imaged pedestrian based on the detected direction of travel of the vehicle, the detected ground contact position of the imaged pedestrian, the detected orientation of the imaged pedestrian, and a set possible collision area.

10. Capture images of the vehicle's surroundings, Detecting a traveling direction of the vehicle; Detecting a landing position of the imaged pedestrian; Detecting the orientation of the captured pedestrian; setting a collision possibility area having a shape and size according to the environment when the pedestrian is imaged; A program that causes an information processing device to determine the possibility of a collision between the vehicle and the imaged pedestrian based on the detected direction of travel of the vehicle, the detected ground position of the imaged pedestrian, the detected orientation of the imaged pedestrian, and the set possible collision area.