Road condition detection system and detection method

Abstract

A road condition detection method and detection system. The detection method comprises the following steps: separately mounting, on a vehicle, a high-position camera having a large included angle between a line of sight thereof and a road surface and a low-position camera having a small included angle between a line of sight thereof and the road surface; the high-position camera and the low-position camera synchronously photographing a detection target on the road surface, respectively obtaining a first image and a second image of the detection target, and transmitting same to an on-board computer; and the on-board computer comparing and analyzing the first image and the second image, and determining a road condition on the basis of the comparison and analysis result. The detection method simulates a human behavioral pattern of observing an object from different angles to determine its authenticity; and by arranging a high-position camera and a low-position camera on a vehicle, road condition detection can be implemented by comparing different image features obtained from different vertical line-of-sight angles, thereby effectively improving the detection capability of existing on-board environment perception systems for road defects, and providing more reliable support for autonomous driving and advanced driver assistance system technologies.

Description

A road condition detection system and detection method Technical Field

[0001] This invention relates to the field of vehicle-mounted environmental perception technology, specifically to a road condition detection system and detection method. Background Technology

[0002] With the rapid development of the automotive industry, intelligent driving is receiving increasing attention. During vehicle operation, the complexity of the road environment poses numerous challenges to the safety of intelligent driving. Among these, road obstacles and potholes are common safety hazards that can lead to vehicle damage or even traffic accidents.

[0003] Currently, video systems for detecting car roadblocks and craters are horizontally arranged. However, obstacles and craters often require vertical comparison and calculation. Vertically arranged video systems have many advantages over horizontally arranged video systems in terms of comparison and calculation.

[0004] In real life, people determine whether a pattern on the road ahead is an obstacle or a pitfall by either walking a few steps forward to observe the changes in the pattern, or by bending down or standing on tiptoe to see it. This illustrates that the best way for the human eye to judge whether an obstacle or pitfall is real or an illusion created by a flat pattern is to observe it from different vertical angles. The same method can be used to determine if the road ahead is a dead end.

[0005] Therefore, the development of a road condition detection system and method that is economical and has a simple analysis process has become a demand. Summary of the Invention

[0006] To address the problems existing in the prior art, the purpose of this invention is to provide a road condition detection system and method that can accurately acquire various road condition information.

[0007] To achieve the above objectives, the present invention adopts the following technical solution:

[0008] A road condition detection method includes the following steps:

[0009] A high-position camera with a large angle between its line of sight and the road surface and a low-position camera with a small angle between their line of sight and the road surface are installed in the car.

[0010] The high-position camera and the low-position camera simultaneously capture images of the detected targets on the road surface, and respectively obtain the first and second images of the detected targets and transmit them to the vehicle-mounted computer.

[0011] The onboard computer compares and analyzes the first and second images, and adds more criteria to the current onboard environmental perception system based on the comparison and analysis results, thereby improving the level of road condition judgment.

[0012] Furthermore, the comparative analysis includes determining that the target is a roadblock or a ditch when the onboard computer detects that the ratio of the longitudinal length of the target in the first image to the height H1 in the second image is not equal to the H1 / H2 of the surrounding road pattern.

[0013] Furthermore, the comparative analysis includes the following: when a certain pattern in the first image is not displayed at the corresponding position in the second image, and the H1 / H2 ratio of the pattern at or below that position in the second image is not equal to the H1 / H2 of the surrounding road pattern, then it is determined that the corresponding position in the second image is obscured by a protruding object on the target being detected, and the protruding object is the upper part of the road barrier.

[0014] Furthermore, the comparative analysis includes determining that the first feature is a pit when the first feature in the first image is not displayed at the corresponding position in the second image, and the H1 / H2 ratio of the pattern at or below that position in the second image is equal to the H1 / H2 of the surrounding road pattern.

[0015] Furthermore, the comparative analysis includes determining that the target is a downhill road when the detected road is a road ahead and the first longitudinal length of the road ahead in the first image is much greater than the second longitudinal length in the second image.

[0016] Furthermore, the comparative analysis includes determining that the target is a dead-end road when the target is the road ahead and the first longitudinal length of the road ahead in the first image is equal to the second longitudinal length in the second image.

[0017] Furthermore, after obtaining the comparative analysis results, the onboard computer will interact with the detected road condition information and the vehicle's autonomous driving or assisted driving system to achieve automatic obstacle avoidance or remind the driver to take appropriate measures.

[0018] A road condition detection system includes a high-position camera, a low-position camera, and an onboard computer;

[0019] High-position and low-position cameras are installed on the car at positions that form different vertical views with the road, and are used to simultaneously capture the detected targets on the road surface, respectively obtaining the first and second images of the detected targets and transmitting them to the on-board computer.

[0020] The onboard computer is used to compare and analyze the first and second images, and to determine the road conditions based on the comparison and analysis results.

[0021] Furthermore, a high-position camera is installed on the roof of the car, and a low-position camera is installed at the front of the car.

[0022] In summary, the present invention has the following advantages:

[0023] This invention simulates the human behavior of observing objects from different angles to determine their authenticity. By arranging high-position cameras, low-position cameras, and an onboard computer on a car, a rough detection of road conditions can be completed through simple analysis. This can effectively improve the detection level of the onboard environmental perception system and provide more reliable support for autonomous driving and assisted driving technologies. Attached Figure Description

[0024] Figure 1 is a schematic diagram of using a road condition detection system to detect road obstacles.

[0025] Figure 2 is a schematic diagram comparing the longitudinal length of the roadblock in the first and second images.

[0026] Figure 3 is a schematic diagram of using a road condition detection system to detect pits.

[0027] Figure 4 is a schematic diagram of using a road condition detection system to detect downhill roads.

[0028] Figure 5 is a schematic diagram of using a road condition detection system to detect dead-end roads.

[0029] In the picture:

[0030] 11 - High-position camera; 12 - Low-position camera;

[0031] 21-Roadblock, 22-Pit, 23-Downhill road, 24-Dead end of road;

[0032] 31 - First feature, 32 - Second feature. Detailed Implementation

[0033] The present invention will now be described in further detail.

[0034] As shown in Figures 1 and 3-5, a road condition detection system includes a high-position camera 11, a low-position camera 12, and an on-board computer.

[0035] The high-position camera 11 and the low-position camera 12 are respectively installed at different heights on the central axis of the car. Preferably, the high-position camera 11 is installed on the top of the car and the low-position camera 12 is installed at the front of the car. The high-position camera 11 is positioned further back than the low-position camera 12.

[0036] The high-position camera 11 and the low-position camera 12 are used to simultaneously capture the detection target on the road surface, respectively obtain the first image and the second image of the detection target, and transmit them to the vehicle computer.

[0037] The onboard computer is used to receive the first image and the second image, and to compare and analyze the first image and the second image. Based on the comparison and analysis results, it judges the road conditions, including analyzing whether the detected target is a roadblock 21, a pit 22, a downhill road 23, or a dead end 24.

[0038] A road condition detection method includes the following steps:

[0039] A high-position camera 11 and a low-position camera 12 are installed at different height positions on the central axis of the car.

[0040] The high-position camera 11 and the low-position camera 12 simultaneously capture images of the detection target on the road surface, respectively obtaining the first image and the second image of the detection target and transmitting them to the vehicle computer.

[0041] The onboard computer compares and analyzes the first and second images, and judges the road conditions based on the comparison and analysis results, including road obstacles 21, pits 22, downhill roads 23, or dead-end roads 24.

[0042] In one embodiment, as shown in Figures 1, 2, and 3, the comparison analysis includes the vehicle-mounted computer acquiring the first longitudinal length H1 of the detected target in the first image (as shown in the left image of Figure 2) and the second longitudinal length H2 in the second image (as shown in the right image of Figure 2). When the ratio between the first longitudinal length H1 and the second longitudinal length H2 exceeds the reference value range, the detected target is determined to be a roadblock 21 or a pit 22. The reference value is acquired by simultaneously capturing planar targets on the road surface using the high-position camera 11 and the low-position camera 12. The vehicle-mounted computer calculates the H1 / H2 value for each horizontal target on the road, and then uses a normal distribution function to calculate the reference value range for that horizontal direction and stores it in the vehicle-mounted computer.

[0043] Specifically, since the longitudinal lengths of planar targets and three-dimensional targets are different in the images captured by the high-position camera 11 or the low-position camera 12, the ratio of the longitudinal length of planar targets in the images captured by the high-position camera 11 and the low-position camera 12 is also different from that of three-dimensional targets. Therefore, the vehicle computer calculates the H1 / H2 value of each horizontal target in the road, and then uses the normal distribution function to calculate the reference value range of the horizontal direction. When the measured target value exceeds the reference value range, it indicates that the detected target is a three-dimensional target, and it can be inferred that the detected target is a roadblock 21 or a pit 22.

[0044] When the high-position camera 11 can capture the first feature 31 behind the roadblock 21 or inside the pit 22, the first image includes the image of the target and the image of the first feature 31 adjacent to and above the target image. Sometimes it also includes the image of the second feature 32 above the image of the first feature 31, as shown in the left image of Figure 2. Due to the obstruction of the roadblock 21 or the pit 22, the low-position camera 12 cannot capture the first feature 31 behind the roadblock 21 or inside the pit 22. It can only capture the second feature 32 behind the first feature 31 (in the case of roadblock 21) or the second feature 32 above the first feature 31 (in the case of pit 22), as shown in the right image of Figure 2. That is, in the second image, the image of the first feature 31 in the first image is missing between the image of the target and the image of the second feature 32. At this time, it can be inferred that the target is the roadblock 21 or the pit 22.

[0045] In another embodiment, the comparison analysis includes determining that when a pattern in the first image is not displayed at the corresponding position in the second image, and the H1 / H2 ratio of the pattern at or below that position in the second image is not equal to the H1 / H2 of the surrounding road pattern, the corresponding position in the second image is determined to be obscured by a protruding object on the target being detected, and the protruding object is the upper part of the roadblock 21.

[0046] As shown in Figures 1 and 2, when the first image contains the image of the target being detected and the image of the first feature 31 adjacent to the target being detected, and the second image contains the image of the target being detected and the image of the second feature 32 adjacent to the target being detected, and the image of the first feature 31 is different from the image of the second feature 32, and the first feature 31 is blocked by a protruding object on the target being detected, the high-position camera 11 can capture the first feature 31, while the low-position camera 12 cannot capture the first feature 31. At this time, it is determined that the protruding object is the top part of the roadblock 21.

[0047] In another embodiment, the comparison analysis includes determining that the first feature 31 is a pit 22 when the first feature 31 in the first image is not displayed at the corresponding position in the second image, and the H1 / H2 ratio of the pattern at or below the location in the second image is equal to the H1 / H2 of the surrounding road pattern.

[0048] As shown in Figure 3, when the first image contains the image of the target being detected and the image of the first feature 31 above the target being detected, and the second image contains the image of the target being detected and the image of the second feature 32 above the target being detected, and the image of the first feature 31 is different from the image of the second feature 32, and the H1 / H2 ratio of the pattern at the corresponding position or below the position in the second image is equal to the H1 / H2 of the surrounding road pattern, it indicates that the target being detected is a planar road rather than a three-dimensional target. At this time, the first feature 31 is determined to be the pit 22.

[0049] When the first feature 31 and the second feature 32 are similar, existing technologies can be used to distinguish them to improve the accuracy of the distinction, which will not be elaborated here.

[0050] In another embodiment, as shown in FIG4, the comparison analysis includes determining that the target is downhill when the first longitudinal length of the target in the first image is much greater than the second longitudinal length in the second image.

[0051] Specifically, due to the obstruction of the downhill road 23, the low-position camera 12 can only capture a small portion of the front end of the downhill road 23, therefore, the second longitudinal length in the second image is shorter. Since the high-position camera 11 is installed higher than the low-position camera 12, it can capture a greater distance of the downhill road 23, resulting in a longer first longitudinal length in the first image, which is typically much greater than the second longitudinal length in the second image. Therefore, by detecting the significant difference in the longitudinal length of the target in the first and second images, it can be inferred that the target is the downhill road 23.

[0052] In another embodiment, as shown in FIG5, the comparison analysis includes determining that the target is a dead-end road 24 when the first longitudinal length of the detected target in the first image is equal to the second longitudinal length in the second image.

[0053] Specifically, when there is a dead end 24 in front of the car, both the high-position camera 11 and the low-position camera 12 can only capture the end of the dead end 24. In the first image and the second image, the longitudinal length of the target is the same. Therefore, it can be inferred that the target is the dead end 24 by the fact that the first longitudinal length of the target in the first image is basically equal to the second longitudinal length in the second image.

[0054] The onboard computer can also calculate information such as the height of the roadblock 21 by using trigonometric functions based on the feature differences of the detected target in the first and second images.

[0055] In real-world driving scenarios, detecting roadblocks 21, pits 22, downhill slopes 23, or dead-end roads 24 often requires observation and analysis from different vertical perspectives to accurately distinguish between real roadblocks 21, pits 22, downhill slopes 23, or dead-end roads 24 and illusions created by planar patterns. This invention proposes a dual-camera image comparison method based on different vertical perspectives for road condition detection in automotive onboard environmental perception systems. By simulating human behavior of observing objects from different angles to determine their authenticity, only a high-position camera 11, a low-position camera 12, and an onboard computer need to be deployed on the vehicle. Simple analysis is sufficient to roughly detect road conditions, effectively improving the onboard environmental perception system's ability to detect road conditions and providing more reliable support for autonomous driving and assisted driving technologies. Compared to deploying lidar, the detection system provided by this invention is more economical, and the detection method is more reliable.

[0056] The above embodiments are preferred embodiments of the present invention, but the embodiments of the present invention are not limited to the above embodiments. Any changes, modifications, substitutions, combinations, or simplifications made without departing from the spirit and principle of the present invention shall be considered equivalent substitutions and shall be included within the protection scope of the present invention.

Claims

1. A road condition detection method, characterized in that: Includes the following steps, A high-position camera with a large angle between its line of sight and the road surface, and a low-position camera with a small angle between its line of sight and the road surface, are installed in the car. The high-position camera and the low-position camera simultaneously capture images of the detected targets on the road surface, and respectively obtain the first and second images of the detected targets and transmit them to the vehicle-mounted computer. The onboard computer compares and analyzes the first and second images, and adds more criteria to the current onboard environmental perception system based on the comparison and analysis results, thereby improving the level of road condition judgment.

2. The road condition detection method according to claim 1, characterized in that: The comparative analysis includes determining whether the target is a roadblock or a ditch when the onboard computer detects that the ratio of the longitudinal length of the target in the first image to the height H1 in the second image is not equal to the H1 / H2 of the surrounding road pattern.

3. The road condition detection method according to claim 2, characterized in that: The comparative analysis includes determining that when a certain pattern in the first image is not displayed at the corresponding position in the second image, and the H1 / H2 ratio of the pattern at or below that position in the second image is not equal to the H1 / H2 of the surrounding road pattern, the corresponding position in the second image is determined to be obscured by a protruding object on the target being detected, and the protruding object is the upper part of the road barrier.

4. The road condition detection method according to claim 2, characterized in that: The comparative analysis includes determining that the first feature is a pit when the first feature in the first image is not displayed at the corresponding position in the second image, and the H1 / H2 ratio of the pattern at or below that position in the second image is equal to the H1 / H2 of the surrounding road pattern.

5. The road condition detection method according to claim 1, characterized in that: The comparative analysis includes determining that the target is a downhill road when the detected road is a road ahead and the first longitudinal length of the road ahead in the first image is much greater than the second longitudinal length in the second image.

6. The road condition detection method according to claim 1, characterized in that: The comparative analysis includes determining that the target is a dead-end road when the target is the road ahead and the first longitudinal length of the road ahead in the first image is equal to the second longitudinal length in the second image.

7. The road condition detection method according to claim 1, characterized in that: After obtaining the comparison and analysis results, the onboard computer will interact with the detected road condition information and the vehicle's autonomous driving or assisted driving system to achieve automatic obstacle avoidance or remind the driver to take appropriate measures.

8. A road condition detection system, characterized in that: This includes high-position cameras, low-position cameras, and in-vehicle computers; High-position and low-position cameras are installed on the car at positions that form different vertical views with the road, and are used to simultaneously capture the detected targets on the road surface, respectively obtaining the first and second images of the detected targets and transmitting them to the on-board computer. The vehicle-mounted computer is used to compare and analyze the first image and the second image according to the method of any one of claims 1-7, and to determine the road conditions based on the comparison and analysis results.

9. The road condition detection system according to claim 8, characterized in that: The high-position camera is installed on the roof of the car, and the low-position camera is installed at the front of the car.

Images