Image processing system, image processing method, and computer program

Abstract

To provide an image processing system capable of preventing a contact between an obstacle and a mobile body regardless of a state of the mobile body.SOLUTION: An image processing system includes: obstacle information acquisition means for detecting the position and height of an obstacle in a travel direction of a mobile body; ground height detection means for detecting a ground height being the height from a ground surface in the travel direction of the mobile body to the mobile body; contact determination means for determining whether the mobile body is to come into contact with the obstacle in the travel direction of the mobile body based on the position and height of the obstacle and the ground height; and warning issuance means for issuing warning when it is determined that the contact is to occur.SELECTED DRAWING: Figure 3

Description

Technical Field

[0001] The present invention relates to an image processing system, an image processing method, a computer program, etc. for preventing contact with an obstacle.

Background Art

[0002] In recent years, as a parking support system, there is a system that displays an overhead image as seen from above the vehicle or an image of the rear of the vehicle on a monitor inside the vehicle. Also, Patent Document 1 describes a technique for suppressing an impact when a wheel contacts a parking stop by performing parking control according to the distance between the parked vehicle and the vehicle when parking.

Prior Art Documents

Patent Documents

[0003]

Patent Document 1

Summary of the Invention

Problems to be Solved by the Invention

[0004] However, in the above-described technology, since vehicle control is performed only considering the height of the parking stop, depending on the height of the parking stop and the state of the vehicle, there is a possibility that the parking stop and the vehicle body may come into contact. For example, when the load of the vehicle is large, the vehicle body sinks, so the possibility of contact between the parking stop and the vehicle body increases.

[0005] Therefore, one object of the present invention is to provide an image processing system that can prevent contact between a moving object and an obstacle regardless of the state of the moving object.

Means for Solving the Problems

[0006] To achieve the above object, an image processing system according to one aspect of the present invention an obstacle information acquisition means for detecting the position and height of an obstacle in the traveling direction of a moving object; A ground height detection means for detecting the ground height, which is the height from the ground in the direction of travel of the moving body to the moving body, A contact determination means that determines whether the moving body and the obstacle come into contact in the direction of the moving body's movement, based on the position and height of the obstacle and the ground clearance. The device is characterized by having a warning output means that outputs a warning when contact is determined to have occurred by the contact determination means. [Effects of the Invention]

[0007] According to the present invention, an image processing system can be realized that can prevent contact between an obstacle and a moving object, regardless of the state of the moving object. [Brief explanation of the drawing]

[0008] [Figure 1] This diagram illustrates the positional relationship between the camera unit and the vehicle as a mobile body in an embodiment of the present invention. [Figure 2] This is a functional block diagram illustrating the configuration of the image processing system in the embodiment. [Figure 3] This is a flowchart illustrating the series of operations of the integrated processing unit 40 in the embodiment. [Figure 4] This figure shows an example of how to obtain the ground clearance h in the direction of travel when vehicle 1 is moving backward. [Figure 5] This figure shows an example of a warning display superimposed on an overhead image displayed on the display unit 50 of the embodiment. [Figure 6] This figure shows an example of a warning display superimposed on an image of the direction of travel displayed on the display unit 50 of the embodiment. [Figure 7] This figure shows another example of a warning display superimposed on an overhead image displayed on the display unit 50 of the embodiment. [Figure 8] This figure shows another example of a warning display superimposed on the image of the direction of travel displayed on the display unit 50 of the embodiment. [Figure 9] (A) and (B) are diagrams showing vehicle 1 in a parking position, viewed from the rear of the vehicle. [Modes for carrying out the invention]

[0009] Embodiments of the present invention will be described below with reference to the drawings. However, the present invention is not limited to the following embodiments. In each drawing, the same reference numeral is used for the same member or element, and redundant explanations are omitted or simplified.

[0010] Examples of the present invention will be explained using Figures 1 to 9. Figure 1 is a diagram illustrating the positional relationship between the camera unit and the vehicle as a mobile body in the embodiment.

[0011] In this embodiment, as shown in Figure 1, camera units 11, 12, 13, and 14 are installed on the front, right side, rear, and left side of a mobile vehicle (mobile body), such as an automobile vehicle 1. Although this embodiment has four camera units, the number of camera units is not limited to four; at least one camera unit is sufficient.

[0012] Furthermore, the camera units 11 to 14 are set to capture images in front of, to the right, to the left, and to the rear of the moving vehicle 1, respectively. In this embodiment, camera units 11 to 14 have substantially the same configuration, and each includes an image sensor for capturing an optical image and an optical system for forming an optical image on the light-receiving surface of the image sensor.

[0013] For example, the optical axes of the optical systems of camera units 11 and 13 are installed so as to be approximately horizontal when the vehicle 1 is horizontal, while the optical axes of the optical systems of camera units 12 and 14 are installed so as to be slightly downward from horizontal or directly downward. Furthermore, the optical systems of the camera units 11 to 14 used in this embodiment are equipped with fisheye lenses and wide-angle lenses that are capable of capturing a wide area of ​​the surroundings.

[0014] Next, the configuration of the image processing system in this embodiment will be explained using Figure 2. FIG. 2 is a functional block diagram for explaining the configuration of the image processing system in the embodiment. Note that some of the functional blocks shown in FIG. 2 are realized by causing a CPU 43 as a computer included in the image processing system 100 to execute a computer program stored in a memory 44 as a storage medium. However, some or all of them may be realized by hardware. As the hardware, a dedicated circuit (ASIC), a processor (reconfigurable processor, DSP), or the like can be used.

[0015] Also, each of the functional blocks shown in FIG. 2 does not have to be built in the same housing, and may be constituted by separate devices connected to each other via signal paths. In FIG. 2, the image processing system 100 is mounted on a vehicle 1 as a moving body, and the camera units 11 to 14 are each composed of an imaging unit 21 to 24 and a camera processing unit 31 to 34.

[0016] The imaging units 21 to 24 each have lenses 21a to 24a and imaging elements 21b to 24b such as a CMOS image sensor or a CCD image sensor. The imaging elements 21b to 24d are so-called imaging surface phase difference type sensors. That is, all pixels on the imaging surface are each composed of a pair of light receiving elements, and at each pixel, a pair of optical images with different parallaxes that have passed through different pupil regions by a microlens are photoelectrically converted by the pair of light receiving elements to generate a pair of imaging signals.

[0017] The lenses 21a to 24a as optical systems are each composed of one or more optical lenses, and form optical images on the light receiving surfaces of the imaging elements 21b to 24b. The imaging elements 21b to 24b function as imaging means, and photoelectrically convert the optical images to output imaging signals. For example, RGB color filters are arranged for each pixel on the light receiving surfaces of the imaging elements 21b to 24b. The RGB color filters are, for example, in a Bayer array.

[0018] Therefore, the image sensor is configured to output signals such as R, G, R, G sequentially from a given row according to the Bayer array, and signals such as G, B, G, B sequentially from the adjacent row. Components 31-34 are camera processing units, which process the imaging signals output from imaging units 21-24, respectively. Note that in Figure 2, the details of the imaging unit 24 and camera processing unit 34, as well as their wiring, are omitted for convenience.

[0019] Each of the camera processing units 31 to 34 has an image processing unit 31a to 34a. Alternatively, part or all of the camera processing unit 31 may be performed by the stacked signal processing units within the image sensors 21b to 24b. The image processing units 31a to 34a each process the imaging signals output from the imaging units 21 to 24.

[0020] Specifically, the image processing units 31a to 34a each debayer the image data input from the imaging units 21 to 24 according to the Bayer array, and convert it into RGB raster image data. Furthermore, they perform various correction processes such as white balance adjustment, gain offset adjustment, gamma processing, color matrix processing, and lossless compression. However, it is desirable not to perform lossy compression.

[0021] Furthermore, as mentioned above, the image sensors 21b to 24b are image-plane phase-difference type image sensors, and the image processing units 31a to 34a create a defocus map using known distance measuring means and send it, for example, to the integrated processing unit 40, where it is stored in the buffer memory 42.

[0022] Furthermore, the camera processing units 31-34 contain a CPU (computer) and memory (storage medium) that stores computer programs. The CPU is configured to control each part of the camera processing units 31-34 by executing the computer programs stored in memory.

[0023] In this embodiment, the image processing units 31a to 34a use hardware such as dedicated circuits (ASICs) or processors (reconfigurable processors, DSPs). This enables faster image recognition in the high-resolution region. The image processing units 31a to 34a may also have a distortion correction function to correct the distortion of the lenses 21a to 24a.

[0024] Furthermore, some or all of the internal functional blocks of the camera processing units 31-34 may be implemented by having the CPU execute computer programs stored in memory; however, in that case, it is desirable to increase the processing speed of the CPU.

[0025] 40 is an integrated processing unit, comprising an SOC (System On Chip) / FPGA (Field Programmable Gate Array) 41, buffer memory 42, a CPU 43 as a computer, and memory 44 as a storage medium. The CPU 43 performs various controls on the entire image processing system 100 by executing computer programs stored in memory 44. In this embodiment, the integrated processing unit 40 is housed in a separate enclosure from the camera unit.

[0026] The SOC / FPGA 41 includes an image acquisition unit 41a, which is a means for receiving output images from the camera processing units 31-34; a bollard information acquisition unit 41b, which is a means for detecting bollard information; and a display image creation unit 41c, which is a means for creating a display image. The image acquisition unit 41a acquires image signals from the camera processing units 31 to 34, outputs the acquired image signals to the bollard information acquisition unit 41b, and also stores them in the buffer memory 42.

[0027] The bollard information acquisition unit 41b detects the presence of obstacles such as bollards by applying a predetermined algorithm to the image in the direction of travel from among multiple image signals input from the image acquisition unit 41a, and detects the size and position information of the bollard (obstacle) on a pixel-by-pixel basis. In other words, the bollard information acquisition unit 41b functions as an obstacle information acquisition unit that detects the position and height of obstacles in the direction of travel of a moving object.

[0028] As a predetermined algorithm, a dataset containing at least images of bollards may be input, and a pre-trained learning model (inference model) may be obtained that has been trained using deep learning to detect the position and height of obstacles in the direction of travel of a moving object from the input images. Then, the position and height of obstacles in the direction of travel of a moving object may be detected from the input images using this learning model (inference model). Alternatively, a configuration may be used to detect bollards based on various known image features such as SIFT, HOG, SURF, ORB, etc.

[0029] Furthermore, the wheel stop information acquisition unit 41b acquires the distance to the wheel stop (obstacle) based on the position of the detected wheel stop (obstacle) and the defocus map stored in the buffer memory 42. It is also desirable to avoid detecting wheel stops in other parking spaces when detecting a wheel stop. For example, the wheel stop detection process may be limited to the image area corresponding to the parking space of vehicle 1 detected by a known parking space detection technique.

[0030] Alternatively, the system may be configured to perform the wheel stop detection process only when the direction of travel is approximately straight relative to the vehicle body of vehicle 1, based on information from the driving control unit (ECU) 60. Furthermore, the means for detecting the distance to the wheel stop may be a stereo camera, or a combination of an imaging camera and a sonar or sensor for acquiring distance information. In the case of a stereo camera, the wheel stop information acquisition unit 41b may perform parallax calculation processing to calculate the distance to the wheel stop.

[0031] The display image creation unit 41c generates a display image to be displayed on the display unit 50 from the images stored in the buffer memory 42. For example, it generates and outputs a display image of the vehicle 1 in the direction of travel, or an overhead view image created by combining multiple images. The display image creation unit 41c also generates an image with warning displays superimposed on the image stored in the buffer memory 42 based on instructions from the CPU 43.

[0032] The display unit 50 displays the display image generated by the display image creation unit 41c. The display unit 50 is composed of a display capable of displaying any image and is installed, for example, around the control panel near the center in the vehicle width direction in front of the driver's seat of the vehicle 1.

[0033] The ECU 60 is a unit that incorporates a computer and memory for comprehensively controlling the drive and direction of the vehicle 1. The ECU 60 is configured to output vehicle control signals to the integrated processing unit 40. The vehicle control signals output by the ECU 60 include information about the vehicle's driving (movement state), such as driving speed, driving direction, shift lever, shift gear, turn signal status, and the vehicle's orientation as determined by a geomagnetic sensor.

[0034] Figure 3 is a flowchart illustrating a series of operations of the integrated processing unit 40 in the embodiment. Each processing step from step S101 to step S105 in Figure 3 is performed sequentially by the CPU 43 of the integrated processing unit 40 executing a computer program in the memory 44.

[0035] In this embodiment, the processing steps from step S101 to step S105 are executed each time the image acquisition unit 41a periodically acquires an image from the camera processing units 31 to 34 while the power supply of the image processing system 100 is ON. Alternatively, the steps may be executed periodically when the vehicle 1 is traveling at a predetermined speed (e.g., 10 km / h) or less, or when a parking space for the vehicle 1 is detected (acquired) using known parking space detection technology.

[0036] In step S101, the CPU 43 refers to the buffer memory 42 and determines whether the wheel stop information acquisition unit 41b has detected a wheel stop. If the CPU 43 determines that a wheel stop has been detected, it proceeds to step S102; otherwise, it terminates the flow shown in Figure 3.

[0037] In step S102, the CPU 43 refers to the buffer memory 42 and obtains the distance (position) to the wheel stops 2a and 2b, as well as the height of the wheel stops, as detected by the wheel stop information acquisition unit 41b, for example, as shown in Figures 4 and 5. In other words, step S102 functions as an obstacle information acquisition step (obstacle information acquisition means) that detects the position and height of obstacles in the direction of travel of the moving object.

[0038] In step S103, the CPU 43 calculates the ground clearance of the vehicle 1 in the direction of travel based on the information acquired in step S102. In other words, step S103 functions as a ground clearance detection step (ground clearance detection means) that detects the ground clearance, which is the height from the ground to the moving body in the direction of travel of the moving body. An example of the calculation method will be explained using Figure 4. Figure 4 shows an example of how to obtain the ground clearance h in the direction of travel when vehicle 1 is moving backward.

[0039] Here, x is the distance from the camera unit 13 to the bottom of the wheel stop 2a, θ is the angle between the horizontal direction of the camera unit 13 and the bottom of the wheel stop 2a, and y is the distance from the horizontal direction of the camera unit 13 to the ground. Also, z is the distance from the horizontal direction of the optical axis of the camera unit 13 to the bottom of the vehicle 1, and h is the ground clearance in the direction of travel of the vehicle 1.

[0040] Here, the angle of the optical axis of the camera unit 13 with respect to the horizontal direction of the ground can be determined, for example, by a tilt sensor inside the camera unit (not shown). On the other hand, the angle from the optical axis of the camera unit 13 to the bottom of the bollard 2a can be determined from the optical characteristics of the camera unit 13 and the acquired image, with a negative value if above the optical axis and a positive value if below.

[0041] The angle θ is calculated by summing these values. Since the distance x can be obtained from the results of the bollard information acquisition unit 41b, the distance y can be calculated using the formula y = x × sinθ. Furthermore, since the camera unit 13 is generally fixed to the vehicle body, the distance z is constant and known. Therefore, the ground clearance h of the vehicle 1 can be calculated using h=yz. In this way, ground clearance can be detected based on the mounting position of the imaging unit on the moving object and the position of obstacles in the image of the imaging unit.

[0042] The above explanation uses Figure 4 to illustrate an example of how to determine the ground clearance of vehicle 1 in the direction of travel in step S103. However, the ground clearance may also be calculated based on the suspension state of vehicle 1 or by attaching a distance measuring sensor to the underside of vehicle 1. That is, a sensor that measures the distance to the ground in the direction of travel of the moving object may be provided, and the ground clearance may be detected based on the distance to the ground obtained by the above sensor. Furthermore, although the explanation used the case where the ground clearance is calculated using wheel chocks 2a, a configuration using wheel chocks 2b may also be used.

[0043] In step S104, the CPU 43 compares the positions and heights of the wheel stops 2a and 2b obtained in step S102 with the ground clearance in the direction of travel of the vehicle 1 calculated in step S103, and determines whether the ground clearance is lower than the height of the wheel stops. In other words, step S104 functions as a contact determination step (contact determination means) that determines whether the moving body and the obstacle will come into contact in the direction of travel of the moving body, based on the position and height of the obstacle and the ground clearance. It then determines that contact will occur if the height of the obstacle is higher than the ground clearance and is located in the direction of travel of the moving body.

[0044] At this time, the higher of the two heights, 2a and 2b, is used for the wheel stop. If the CPU 43 determines that the ground clearance is higher than the height of the wheel stop, it terminates this process. On the other hand, if the CPU 43 determines that the ground clearance is less than or equal to the height of the wheel stop, it proceeds to step S105.

[0045] In step S105, the CPU 43 executes a process to issue a warning to the driver. That is, step S105 is a step to output a warning if it is determined in step S104 that contact has occurred. Specifically, in this embodiment, the CPU 43 outputs an instruction to superimpose a warning image onto the display image generated by the display image creation unit 41c. Examples of warning images are explained using Figures 5 to 8.

[0046] Figure 5 shows an example of a warning display superimposed on an overhead image displayed on the display unit 50 of the embodiment. When generating an overhead image as a display image using the display image creation unit 41c, a warning image 51 as shown in Figure 5 can be superimposed to warn the driver. In this way, the display image creation unit 41c may create an overhead image from the image of the imaging unit and then create an image in which a warning display based on the result of the contact detection is superimposed on the overhead image.

[0047] Figure 6 shows an example of a warning display superimposed on the image of the direction of travel displayed on the display unit 50 of the embodiment. In Figure 6, when the image of the direction of travel extracted by the display image creation unit 41c is generated as a display image, the warning image 51 is superimposed to warn the driver.

[0048] Thus, the display image creation unit 41c may create a direction-of-movement image of the moving object from the image captured by the imaging unit, and then create an image by superimposing a warning display based on the contact detection result onto the direction-of-movement image. Furthermore, it is desirable that the warning image 51 in Figures 5 and 6 be superimposed in a position that does not overlap with the wheel stops 2a and 2b.

[0049] Figure 7 shows another example of a warning display superimposed on an overhead image displayed on the display unit 50 of the embodiment. In Figure 7, when the overhead image created by the display image creation unit 41c is generated as a display image, the position 52 just before the vehicle body 1 and the wheel stop are superimposed. This allows the driver to see a position where they can park safely. The position where there is no contact can be calculated by setting it to a position a predetermined distance (for example, 30 cm) before the shorter of the distances to the wheel stops 2a and 2b detected by the wheel stop information acquisition unit 41b.

[0050] At this time, if the installation position of the camera unit 13 is offset in front of or behind the lower rear end of the vehicle, this is also taken into consideration. Then, when the vehicle 1 is reversing, by superimposing the portion where the distance to the wheel stop is, for example, 30 cm from the lower rear end of the vehicle, it is possible to display a parking position where the lower rear end of the vehicle does not come into contact with the wheel stop.

[0051] Figure 8 shows another example of a warning display superimposed on the image of the direction of travel displayed on the display unit 50 of the embodiment. In Figure 8, when the image of the direction of travel extracted by the display image creation unit 41c is generated as a display image, the position 52 where the vehicle body 1 and the wheel stop do not come into contact is superimposed. This makes it possible to show the driver a safe parking position.

[0052] Furthermore, while an example of displaying a warning image superimposed on the display image generated by the display image creation unit 41c and displayed on the display unit was described as a method of warning the driver, this is not limited to this method as long as it is possible to inform the driver of the possibility of contact. For example, a speaker control unit that controls the speaker mounted on the vehicle 1 may be provided and configured to communicate with the image processing system 100, and the CPU 43 may cause the speaker control unit to output a warning sound. In other words, the warning output is not limited to the display of a warning image, but may also include warnings by outputting sound.

[0053] Furthermore, for example, a sensor may be provided to acquire suspension status information, which is information about the state of the suspension. The sensor may detect the ground clearance of the moving body, and if it is determined that the load is too high, an announcement may be made by voice to reduce the load. Alternatively, a combination of the above-mentioned voice warning and image display warning may be used to provide a warning.

[0054] Furthermore, in step S105, the CPU 43 may be configured to output an instruction to the driving control unit (ECU) to apply the brakes. That is, when a warning is output, the CPU 43 may output a deceleration instruction or a stop instruction to the moving object. By activating the automatic braking system in this way, a collision can be avoided even if the driver fails to notice the warning or is slow to respond to it.

[0055] The effects obtained by the operation of each part of this embodiment as described above will be explained with reference to Figure 9. Figures 9(A) and (B) show the state in which vehicle 1 is parked in reverse, as seen from the rear of the vehicle. Figure 9(A) shows the case when the height from the ground in the direction of travel of vehicle 1 to the bottom of the vehicle body (ground clearance) is higher than the wheel stops 2a and 2b.

[0056] In the case shown in Figure 9(A), vehicle 1 can be parked without its body touching the wheel stops 2a and 2b. At this time, the image processing system 100 determines that there is no contact and does not issue a warning to the driver. Therefore, unnecessary warnings can be reduced.

[0057] On the other hand, Figure 9(B) shows the case where the ground clearance of vehicle 1 in the direction of travel is lower than the wheel stops 2a and 2b. In the case shown in Figure 9(B), there is a possibility that the body of vehicle 1 will come into contact with the wheel stops 2a and 2b before the tires of vehicle 1. For example, if vehicle 1 is heavily loaded, the body of vehicle 1 will sink, increasing the possibility of contact between the body of vehicle 1 and the wheel stops 2a and 2b. In such a case, according to the image processing system 100 of this embodiment, the ground clearance of the bottom of vehicle 1 in the direction of travel is calculated and compared with the height of the wheel stops 2a and 2b, determining that contact is likely and issuing a warning to the driver, thereby appropriately warning the driver.

[0058] In the above embodiment, the system is configured to detect wheel stops 2a and 2b in the direction of travel of vehicle 1 and issue a warning if the ground clearance is lower than the wheel stops 2a and 2b. However, the detection target in this embodiment is not limited to wheel stops. For example, it may also be an obstacle other than a wheel stop (such as fallen objects on the road, animals, or people).

[0059] In the above-described embodiments, an example was given in which an image processing system was mounted on a moving object such as a vehicle. However, the moving object in the embodiments is not limited to vehicles such as automobiles, but can be any mobile device that moves on the ground, such as a train or a robot. Furthermore, the image processing system in the embodiments may or may not be mounted on such moving objects. In addition, the configurations of these embodiments can also be applied, for example, when remotely controlling a moving object.

[0060] Although the present invention has been described in detail above based on preferred embodiments, the present invention is not limited to the above embodiments, and various modifications are possible in accordance with the spirit of the present invention, and these modifications are not excluded from the scope of the present invention.

[0061] Furthermore, the present invention may also be realized by supplying a storage medium containing software program code (control program) that realizes the functions of the embodiments described above to an image processing system or the like. It can also be achieved by the computer (or CPU or MPU) of the image processing system or the like reading and executing the computer-readable program code stored on the storage medium. In that case, the program code read from the storage medium itself will realize the function of the embodiment described above, and the storage medium storing that program code will constitute the present invention. [Explanation of Symbols]

[0062] 1: Vehicle 11-14: Camera Unit 21-24: Imaging Department 31-34: Camera Processing Unit 40: Integrated Processing Unit 41a: Image acquisition unit 41b: Vehicle Stop Information Acquisition Unit 41c: Display image creation section 50: Display section 60: Driving Control Unit (ECU) 100: Image Processing System

Claims

1. An obstacle information acquisition means for detecting the position and height of an obstacle in the traveling direction of a moving body; A ground height detection means for detecting the ground height, which is the height from the ground to the moving body in the traveling direction of the moving body; A contact determination means for determining whether or not the moving body and the obstacle come into contact with each other in the traveling direction of the moving body based on the position and height of the obstacle and the ground height; A warning output means for outputting a warning when it is determined by the contact determination means that they come into contact; At least one imaging means, and The ground height detection means detects the ground height based on the position of the imaging means in the moving body and the position of the obstacle in the image of the imaging means An image processing system characterized by the above.

2. The contact determination means determines that they come into contact when the height of the obstacle is higher than the ground height, according to the image processing system described in Claim 1.

3. The image processing system according to Claim 1, further comprising a display image creation means for creating an overhead image from the image of the imaging means and superimposing a warning display based on the result of the contact determination means on the overhead image.

4. The image processing system according to Claim 1, further comprising a display image creation means for creating a traveling direction image regarding the traveling direction of the moving body from the image of the imaging means and superimposing a warning display based on the result of the contact determination means on the traveling direction image.

5. The warning output means includes an audio output means, according to the image processing system described in Claim 1.

6. The warning output means outputs a deceleration instruction to the moving body, according to the image processing system described in Claim 1.

7. The image processing system according to Claim 1, wherein the ground height detection means detects the ground height based on the attachment position of the imaging means to the moving body and the position of the obstacle on the image of the imaging means.

8. further comprising a suspension state information acquisition means for acquiring suspension state information, which is the state information of the suspension of the moving body, The ground height detection means detects the ground height of the moving body based on the suspension state information, according to the image processing system described in Claim 1.

9. The ground height detection means has a sensor for measuring the distance to the ground in the traveling direction of the moving body The image processing system according to claim 1, wherein the ground height is detected based on the distance to the ground acquired by the sensor.

10. The obstacle information acquisition means inputs a data set including an image of the obstacle, and uses a learning model that has been pre-learned to detect the position and height of the obstacle in the traveling direction of the moving body from the input image, and detects the position and height of the obstacle in the traveling direction of the moving body from the image. The image processing system according to claim 1, characterized by the above.

11. The image processing system according to claim 1, wherein the obstacle includes a vehicle stopper.

12. An obstacle information acquisition step of detecting the position and height of an obstacle in the traveling direction of the moving body; A ground height detection step of detecting the ground height, which is the height from the ground in the traveling direction of the moving body to the moving body; A contact determination step of determining whether or not the moving body contacts the obstacle based on the height of the obstacle and the ground height; A warning output step of outputting a warning when it is determined that there is contact in the contact determination step; And at least one imaging step, In the ground height detection step, the ground height is detected based on the position of the imaging step in the moving body and the position of the obstacle in the image of the imaging step. An image processing method characterized by the above.

13. A computer program for controlling each means of the image processing system according to any one of claims 1 to 11 by a computer.

Images