Vehicle front 3D detection system based on monocular camera
By setting a reference grid in front of the monocular camera and using the standard lines of the reference grid to determine image distortion, the computational workload of image processing is simplified, the problem of image inaccuracy caused by radial distortion of the monocular camera is solved, and efficient image reconstruction is achieved.
Patent Information
- Application Number
- CN202311379717.7
- Authority / Receiving Office
- CN · China
- Patent Type
- Patents(China)
- Current Assignee / Owner
- Filing Date
- 2023-10-23
- Publication Date
- 2024-11-05
- Estimated Expiration
- Not applicable · inactive patent
AI Technical Summary
Monocular cameras often suffer from radial distortion when acquiring images, resulting in inaccurate image information. Existing technologies require complex calculations to compensate for the distortion.
A reference grid is set up in front of the monocular camera. By comparing the standard lines of the reference grid with the distorted image, the degree of distortion of the environmental image can be quickly determined. The image is then reconstructed by rotating the platform and the processor, simplifying the computation.
It can quickly identify and repair image distortions, reduce the computational load of image processing, and improve the accuracy and efficiency of image reconstruction.
Smart Images

Figure CN117455759B_ABST
Abstract
Description
Technical Field
[0001] The present invention belongs to the field of intelligent driving technology, and in particular to a three-dimensional detection system for a vehicle ahead based on a monocular camera. Background Art
[0002] At present, intelligent driving is one of the main development directions of vehicle technology. The key to intelligent driving is to accurately and reliably obtain road condition information. Commonly used image acquisition devices include monocular cameras, multi-camera cameras and depth cameras. Among them, monocular cameras are widely used in the fields of three-dimensional detection and imaging due to their simple structure and low cost. For example, CN202010668524.3-A method and device for calculating the position of a target object based on a monocular camera, CN202011383475.5-A pedestrian active collision avoidance system based on a monocular camera System and method, CN202010471623.2-A monocular camera ranging method using laser assisted calibration, CN202010178616.3-A vehicle-pedestrian collision avoidance method based on a visual system, CN201810993791.0-A method for realizing three-dimensional measurement of dynamic points using a monocular camera, CN201921316283.5-A device for generating a three-dimensional image, and other prior arts, all of which utilize a monocular camera to obtain image information, and then generate three-dimensional information of the object based on the image information.
[0003] However, when a monocular camera acquires an image, radial distortion often occurs. There are two common radial distortions: barrel distortion, such as Figure 1 As shown, the other is pincushion distortion, such as Figure 2 In order to improve the accuracy of image information, it is necessary to compensate the image according to the distortion formula, which requires a lot of calculation. Summary of the invention
[0004] The technical problem to be solved by the present invention is to provide a three-dimensional detection system for the front of a vehicle based on a monocular camera, which can eliminate the influence of radial distortion, and at the same time, the image processing process is relatively simple and the calculation amount is low.
[0005] To solve the above problems, the technical solution adopted by the present invention is: a three-dimensional vehicle front detection system based on a monocular camera, comprising a monocular camera, a processor and a memory, wherein the monocular camera is connected to the processor.
[0006] A reference grid is arranged in front of the monocular camera, the reference grid comprises a plurality of mutually parallel first standard lines and a plurality of mutually parallel second standard lines, the first standard lines being perpendicular to the second standard lines;
[0007] The memory is used to store the standard image of the reference network;
[0008] The monocular camera is used to obtain image information in front of the vehicle and transmit the image information to the processor, wherein the image includes an environment image and a reference network image;
[0009] The processor is used to compare the reference network image in the image information with the standard image, determine the degree of distortion of the environment image, reconstruct the environment image, and finally generate a three-dimensional image of the environment in front of the vehicle.
[0010] Furthermore, the monocular camera includes a first monocular camera and a second monocular camera, the first monocular camera is installed on a left-right rotating platform, and the second monocular camera is installed on an up-and-down rotating platform;
[0011] The left-right rotating platform is used to drive the first monocular camera to rotate left-right. The first monocular camera scans the environment in front of the vehicle during the rotation process and obtains a plurality of first image information. Each first image information includes a first environment image and a first reference network image. The first reference network image divides the first environment image into a plurality of first partial images.
[0012] The vertical rotation platform is used to drive the second monocular camera to rotate up and down. The second monocular camera scans the environment in front of the vehicle during the rotation and obtains a plurality of second image information. Each second image information includes a second environment image and a second reference network image. The second reference network image divides the second environment image into a plurality of second partial images.
[0013] The processor compares the first reference network image with the standard image of the reference network, finds the first partial image whose distortion degree meets the requirement in each first environment image, and cuts the first partial image whose distortion degree is less than the set value to obtain multiple first target images;
[0014] The processor compares the second reference network image with the standard image of the reference network, finds the second partial image whose distortion degree meets the requirement in each second environment image, and cuts the second partial image whose distortion degree is less than the set value to obtain a plurality of second target images;
[0015] The processor splices the plurality of first target images and the plurality of second target images to obtain a target environment image for generating a three-dimensional image.
[0016] Further, the first environment images are scaled according to the sizes of the objects in the different first environment images so that the sizes of the objects in all the first environment images are consistent, and then the first target image is acquired;
[0017] The second environment images are scaled according to the sizes of the objects in the different second environment images so that the sizes of the objects in all the second environment images are consistent, and then the second target image is acquired.
[0018] Furthermore, it also includes a laser rangefinder, which is used to measure the distance from the first monocular camera and the second monocular camera to the front object.
[0019] Furthermore, it also includes a shell, the left-right rotating platform and the up-down rotating platform are both installed in the shell, the shell has an open end, and the reference network is arranged in the open end.
[0020] Further, the left-right rotating platform includes a first mounting tube and a support frame, the middle part of the first mounting tube is hinged to the support frame, one end of the first mounting tube is connected to a driving mechanism, and the first monocular camera is installed at the other end of the first mounting tube;
[0021] The up and down rotating platform includes a second mounting cylinder and a push plate, the middle part of the second mounting cylinder is hinged to the support frame, one end of the second mounting cylinder is hinged to the push plate, and the second monocular camera is installed on the other end of the second mounting cylinder; the push plate and the support frame slide in cooperation in the vertical direction, the lower surface of the push plate is provided with an arc-shaped slide groove, the bottom of the slide groove is provided with a first inclined surface, a transmission block is provided on the first mounting cylinder, the upper surface of the transmission block is provided with a second inclined surface, the transmission block is located in the slide groove and the first inclined surface is in contact with the second inclined surface, and a compression spring is provided between the upper surface of the push plate and the top wall of the shell.
[0022] Furthermore, the driving mechanism comprises a linear motor, the linear motor is connected to a telescopic column, and the telescopic column is hinged to the first mounting cylinder.
[0023] Furthermore, a transparent protective cover is provided at the open end of the shell, and the reference network is provided inside the protective cover.
[0024] Furthermore, a positioning ring is provided on the outer side of the protective cover, and a plurality of nozzles are provided on the positioning ring. The nozzles are connected to a driving pump through a pipeline, and the driving pump is connected to an air inlet pipe and a water inlet pipe. Valves are provided on the air inlet pipe and the water inlet pipe, and the water inlet pipe is connected to a water tank.
[0025] Furthermore, the first standard wire and the second standard wire are steel wires with a diameter of 0.1 mm, and a black coating is provided on the surface of the steel wires.
[0026] The beneficial effects of the present invention are as follows: by setting a reference network in front of a monocular camera, when the monocular camera captures an image in front of a vehicle, the reference network will also be captured, so that the image information obtained includes not only an environmental image but also a reference network image, and the reference network image divides the environmental image into multiple parts. Under standard (i.e., undistorted) conditions, the first standard line and the second standard line in the reference network image are straight lines, and after distortion, the first standard line and the second standard line become curves. Therefore, the position, form and degree of distortion of the environmental image can be determined based on the deformation of the first standard line and the second standard line, and then repair can be performed in a targeted manner. By introducing a physical reference network, the present invention can quickly determine the distortion condition of the environmental image without complex calculations, thereby simplifying the amount of calculation for image processing and reducing the load on the processor. BRIEF DESCRIPTION OF THE DRAWINGS
[0027] Figure 1 It is a schematic diagram of barrel distortion;
[0028] Figure 2 This is a schematic diagram of pincushion distortion;
[0029] Figure 3 is a schematic block diagram of the components of the present invention;
[0030] Figure 4 It is a schematic diagram of installing a monocular camera of the present invention.
[0031] Figure 5 yes Figure 4 Schematic cross-sectional view of AA in the figure;
[0032] Figure 6 yes Figure 4 A schematic cross-sectional view of the middle BB;
[0033] Figure 7 It is a bottom view schematic diagram of the push plate and the transmission block;
[0034] Figure 8 It is a schematic diagram of the driving mechanism. DETAILED DESCRIPTION
[0035] The present invention is further described below in conjunction with the accompanying drawings and embodiments.
[0036] The vehicle front three-dimensional detection system based on a monocular camera of the present invention is as follows: Figures 3 to 8 As shown, it includes a monocular camera 1, a processor 2 and a memory 3, and the monocular camera 1 is connected to the processor 2. The hardware of the monocular camera 1, the processor 2 and the memory 3 can adopt the existing technology.
[0037] A reference grid is arranged in front of the monocular camera 1 . The reference grid includes a plurality of mutually parallel first standard lines 5 and a plurality of mutually parallel second standard lines 4 . The first standard lines 5 are perpendicular to the second standard lines 4 .
[0038] The reference grid is a physical grid, which is arranged in front of the lens of the monocular camera 1. When the monocular camera 1 is shooting, the reference grid can be simultaneously captured in the image. Therefore, the first standard line 5 and the second standard line 4 of the reference grid can be used as reference standards, and the distortion of other environmental images can be judged according to the distortion of the first standard line 5 and the second standard line 4.
[0039] The memory 3 is used to store a standard image of the reference network; the standard image refers to a reference network image without distortion, in which the first standard line 5 and the second standard line 4 are straight lines, and multiple first standard lines 5 are parallel to each other, multiple second standard lines 4 are parallel to each other, the second standard lines 4 and the first standard lines 5 are perpendicular to each other, and the second standard lines 4 and the first standard lines 5 form a standard grid.
[0040] The monocular camera 1 is used to obtain image information in front of the vehicle and transmit the image information to the processor 2, and the image includes an environment image and a reference network image.
[0041] Processor 2 is used to compare the reference network image in the image information with the standard image, determine the degree of distortion of the environment image, reconstruct the environment image, and finally generate a three-dimensional image of the environment in front of the vehicle.
[0042] In the standard image, the second standard line 4 and the first standard line 5 are straight lines without deformation. When the image captured by the monocular camera 1 is distorted, the reference network image will also be distorted, that is, the second standard line 4 and the first standard line 5 are bent and deformed. The second standard line 4 and the first standard line 5 are used as references, and according to the deformation position and deformation amplitude of the second standard line 4 and the first standard line 5, the deformation position and deformation amplitude of the environment image can be quickly determined to facilitate reconstruction of the environment image.
[0043] By setting a physical reference network as a reference, the distortion of the environmental image can be quickly and intuitively determined, and then the environmental image can be quickly reconstructed.
[0044] The specific image reconstruction process of the present invention is:
[0045] Two monocular cameras 1 are provided, specifically a first monocular camera and a second monocular camera, the first monocular camera is mounted on a left-right rotating platform 6, and the second monocular camera is mounted on a vertical rotating platform 7. The left-right rotating platform 6 can drive the first monocular camera to rotate left-right, and during the rotation process, the height of the first monocular camera does not change. The vertical rotating platform 7 is used to drive the second monocular camera to rotate up and down, and during the rotation process, the horizontal position of the second monocular camera does not change.
[0046] The left-right rotating platform 6 is used to drive the first monocular camera to rotate left-right. During the rotation process, the first monocular camera scans the environment in front of the vehicle and obtains multiple first image information. Each first image information includes a first environment image and a first reference network image. The first reference network image divides the first environment image into multiple first local images.
[0047] The vertical rotating platform 7 is used to drive the second monocular camera to rotate up and down. During the rotation process, the second monocular camera scans the environment in front of the vehicle and obtains multiple second image information. Each second image information includes a second environment image and a second reference network image. The second reference network image divides the second environment image into multiple second local images.
[0048] Processor 2 compares the first reference network image with the standard image of the reference network, finds out the first local image whose distortion degree meets the requirement in each first environment image, and crops the first local image whose distortion degree is less than the set value to obtain multiple first target images.
[0049] Processor 2 compares the second reference network image with the standard image of the reference network, finds out the second local image whose distortion degree meets the requirement in each second environment image, and crops the second local image whose distortion degree is less than the set value to obtain multiple second target images.
[0050] The processor 2 splices the multiple first target images and the multiple second target images to obtain a target environment image for generating a three-dimensional image.
[0051] The left-right rotating platform 6 and the up-down rotating platform 7 take pictures of the environment in front of the vehicle while rotating, thereby obtaining a plurality of first environment images and second environment images. Figure 1 and Figure 2 It can be seen that, whether it is barrel distortion or pincushion distortion, the parts with the largest deformation are at the edge of the image, and the center of the image is slightly deformed and can be ignored. Therefore, in each first environment image, the first local image with the required distortion is found and cropped to obtain multiple first target images. Similarly, multiple second target images can be obtained, and then the multiple first target images and the multiple second target images are spliced.
[0052] The degree of distortion of the first partial image and the second partial image is determined based on the degree of deformation of the second standard line 4 and the first standard line 5 at the edge of the first partial image and the second partial image. During image reconstruction, there may be situations where two second target images have overlapping parts, two first target images have overlapping parts, and the second target image and the first target image have overlapping parts. When the two images completely overlap, one image is discarded; when the two images partially overlap, the overlapping part of one image is cut off, and the overlapping part of the other image is retained.
[0053] The present invention installs the first monocular camera and the second monocular camera on a rotating platform, which has the following advantages:
[0054] 1. The first monocular camera and the second monocular camera have a wider shooting range, and can more comprehensively obtain larger objects in front of the vehicle, such as a large truck that is located in front of the vehicle and close to the vehicle, and can also obtain road condition information on the left and right front sides of the vehicle, which is conducive to ensuring safety.
[0055] 2. According to the characteristics of barrel distortion or pincushion distortion, the position with the smallest deformation is the center of the image, that is, the first target image and the second target image are both the first partial image and the second partial image located at the center of the image. In the process of the first monocular camera and the second monocular camera rotating and acquiring images, the position of the object in front of the vehicle located at the center of the image is constantly changing, so that different parts of the object are located at different image centers. Therefore, the cropped first target image and the second target image contain more parts of the object, which can more realistically restore the original appearance of the object and improve the accuracy of image reconstruction.
[0056] When the vehicle is moving, the distance between the vehicle and the object in front may change continuously. In the images acquired by the first monocular camera and the second monocular camera at different time points, the size of the same object changes and is not uniform. Therefore, before image cropping, the image should be scaled according to the change in the distance between the camera and the object in front so that the sizes of the same object in different images match, and then cropping and splicing are performed to ensure the accuracy of image reconstruction.
[0057] Specifically, the first environment images are scaled according to the sizes of objects in different first environment images so that the sizes of objects in all first environment images are consistent, and then the first target image is acquired;
[0058] When the same object is close to the first monocular camera, the size of the first image information is larger; when it is far from the first monocular camera, the size of the first image information is smaller. Therefore, when the distance between the object and the first monocular camera changes, the size of the object in the first environment image acquired by the first monocular camera at different times also changes. In order to keep the ratio of each image consistent, the present invention scales the first environment image according to the size of the object on the first environment image, so that the size of the same object in all first environment images is consistent, and then cropping and reconstruction are performed. Since the distance change from the reference network to the first monocular camera is very small and can be ignored, during the scaling process, only the first environment image is scaled, and the first reference network image is not scaled.
[0059] Similarly, the second environment images are scaled according to the sizes of the objects in different second environment images so that the sizes of the objects in all the second environment images are consistent, and then the second target image is acquired.
[0060] The present invention further includes a laser rangefinder 8, which is used to measure the distance between the first monocular camera and the second monocular camera and the object in front. When the distance between the vehicle and the object in front is relatively close, and the time required for the vehicle to reach the object in front is relatively short according to the vehicle speed, an alarm can be sounded to remind the driver to stop the vehicle in time. In addition, since the monocular camera adopts the pinhole imaging principle, when the distance between the monocular camera and the object in front is determined, the actual size of the object can be calculated according to the size of the object in the image and the focal length, so as to more accurately establish a three-dimensional image of the object.
[0061] In order to facilitate the installation of the system as a whole on the vehicle, the present invention further comprises a housing 9, in which the left-right rotating platform 6 and the up-and-down rotating platform 7 are both installed. The housing 9 has an open end, in which the reference network is arranged.
[0062] The housing 9 can be a thin-walled metal housing or a plastic housing as long as the strength meets the requirements. The housing 9 is in the shape of a rectangular parallelepiped and has a rectangular inner cavity. One end of the housing 9 is open and the other end is closed. The lenses of the first monocular camera and the second monocular camera face the open end to facilitate shooting of the external environment. The processor 2 and the memory 3 can also be arranged inside the housing 9. The laser rangefinder 8 is arranged outside the housing 9 and installed separately.
[0063] The left-right rotating platform 6 and the up-down rotating platform 7 can be driven by a driving device respectively. In order to simplify the structure, the left-right rotating platform 6 of the present invention includes a first mounting tube 61 and a support frame 62. The middle part of the first mounting tube 61 is hinged to the support frame 62. One end of the first mounting tube 61 is connected to the driving mechanism 63, and the first monocular camera is installed at the other end of the first mounting tube 61. The up and down rotating platform 7 includes a second mounting cylinder 71 and a push plate 72. The middle part of the second mounting cylinder 71 is hinged to the support frame 62, one end of the second mounting cylinder 71 is hinged to the push plate 72, and the second monocular camera is installed at the other end of the second mounting cylinder 71; the push plate 72 and the support frame 62 slide in the vertical direction, and the lower surface of the push plate 72 is provided with an arc-shaped slide groove, and the bottom of the slide groove is provided with a first inclined surface. A transmission block 64 is provided on the first mounting cylinder 61, and the upper surface of the transmission block 64 is provided with a second inclined surface. The transmission block 64 is located in the slide groove and the first inclined surface is in contact with the second inclined surface. A compression spring 73 is provided between the upper surface of the push plate 72 and the top wall of the shell 9.
[0064] The first mounting tube 61 and the second mounting tube 71 are circular cylinders with open ends, and the first monocular camera and the second monocular camera are respectively mounted on the ports of the first mounting tube 61 and the second mounting tube 71. Since the middle of the first mounting tube 61 is hinged with the support frame 62, the first mounting tube 61 can swing left and right under the drive of the driving mechanism 63, thereby driving the first monocular camera to move left and right, and realizing left and right scanning of the object in front.
[0065] The push plate 72 is horizontally arranged above the transmission block 64 and can slide up and down. One end of the second mounting tube 71 is hinged to the push plate 72 through a connecting rod. The middle part of the second mounting tube 71 is hinged to the support frame 62. When the push plate 72 moves up and down, it can drive the second mounting tube 71 to swing up and down, and then drive the second monocular camera to rotate up and down, so as to realize the up and down scanning of the object in front.
[0066] When the first mounting cylinder 61 swings left and right, the transmission block 64 moves with the first mounting cylinder 61, and its motion track is an arc. Therefore, an arc-shaped slide groove is provided on the lower surface of the push plate 72, and the shape of the slide groove is the same as the shape of the motion track of the transmission block 64, so that the transmission block 64 can slide in the slide groove. The compression spring 73 is always in a compressed state. Under the elastic force of the compression spring 73, the bottom of the slide groove always fits the upper surface of the transmission block 64. Since the bottom of the slide groove and the upper surface of the transmission block 64 are inclined surfaces, when the transmission block 64 moves, it will push the push plate 72 to move upward, or the push plate 72 will move downward under the elastic force of the compression spring 73, thereby driving the second mounting cylinder 71 to swing up and down.
[0067] The present invention only needs one driving mechanism 63 to drive the left-right rotating platform 6 and the up-down rotating platform 7 to move simultaneously, which is helpful to simplify the system structure.
[0068] The driving mechanism 63 specifically includes a linear motor, the linear motor is connected to a telescopic column 65 , and the telescopic column 65 is hinged to the first mounting cylinder 61 .
[0069] The three-dimensional detection system of the present invention can be installed inside the front windshield or on the head of the vehicle. When installed on the head of the vehicle, the three-dimensional detection system is exposed to the external environment, and rain, dust, etc. easily enter the inside of the housing 9, making it inconvenient to clean. Therefore, a transparent protective cover 10 is provided at the open end of the housing 9, and the reference net is provided inside the protective cover 10. The protective cover 10 can be a proportional cover, etc., to play a protective role.
[0070] When the vehicle is driving, dust will adhere to the protective cover 10, causing the images obtained by the first monocular camera and the second monocular camera to be blurred, affecting the accuracy of image reconstruction. Therefore, a positioning ring 11 is provided on the outside of the protective cover 10, and a plurality of nozzles 12 are provided on the positioning ring 11. The nozzles 12 are connected to a driving pump 13 through a pipeline. The driving pump 13 is connected to an air intake pipe 14 and a water intake pipe 15. Valves are provided on the air intake pipe 14 and the water intake pipe 15, and the water intake pipe 15 is connected to a water tank 16. When there is a lot of dust on the protective cover 10, the driving pump 13 can be started. The driving pump 13 first delivers water in the water tank 16 to each nozzle 12, and the nozzle 12 sprays water to the outer wall of the protective cover 10 to clean the protective cover 10. After cleaning, the water supply is stopped, and the driving pump 13 delivers air to the nozzle 12 to dry the moisture on the outer wall of the protective cover 10.
[0071] The first standard line 5 and the second standard line 4 are specifically made of steel wire with a diameter of 0.1 mm. The surface of the steel wire is provided with a black coating, and the coating can be anti-corrosion paint, etc. The first standard line 5 and the second standard line 4 are painted with black water, which will not reflect light during shooting, and a black reference network image is obtained on the image.
[0072] The above description is only a preferred embodiment of the present invention and is not intended to limit the present invention. For those skilled in the art, the present invention may have various modifications and variations. Any modification, equivalent replacement, improvement, etc. made within the spirit and principle of the present invention shall be included in the protection scope of the present invention.
Claims
1. A vehicle front three-dimensional detection system based on a monocular camera, comprising a monocular camera (1), a processor (2) and a memory (3), wherein the monocular camera (1) is connected to the processor (2), and is characterized in that: A reference network is arranged in front of the monocular camera (1), the reference network comprising a plurality of mutually parallel first standard lines (5) and a plurality of mutually parallel second standard lines (4), the first standard lines (5) being perpendicular to the second standard lines (4); The memory (3) is used to store a standard image of a reference network, in which the first standard line (5) and the second standard line (4) are straight lines, and a plurality of first standard lines (5) are parallel to each other, a plurality of second standard lines (4) are parallel to each other, the second standard lines (4) and the first standard lines (5) are perpendicular to each other, and the second standard lines (4) and the first standard lines (5) form a standard grid; The monocular camera (1) is used to obtain image information in front of the vehicle and transmit the image information to a processor (2), wherein the image includes an environment image and a reference network image; The processor (2) is used to compare the reference network image in the image information with the standard image, determine the position, form and degree of distortion of the environment image according to the deformation position and deformation amplitude of the first standard line (5) and the second standard line (4), reconstruct the environment image, and finally generate a three-dimensional image of the environment in front of the vehicle; The monocular camera (1) comprises a first monocular camera and a second monocular camera, wherein the first monocular camera is mounted on a left-right rotating platform (6), and the second monocular camera is mounted on a top-bottom rotating platform (7); The left-right rotating platform (6) is used to drive the first monocular camera to rotate left-right, and the first monocular camera scans the environment in front of the vehicle during the rotation process and obtains a plurality of first image information, each of which includes a first environment image and a first reference network image, and the first reference network image divides the first environment image into a plurality of first partial images; The vertical rotation platform (7) is used to drive the second monocular camera to rotate vertically, and the second monocular camera scans the environment in front of the vehicle during the rotation process and obtains a plurality of second image information, each of which includes a second environment image and a second reference network image, and the second reference network image divides the second environment image into a plurality of second partial images; The processor (2) compares the first reference network image with the standard image of the reference network, finds the first partial image whose distortion degree meets the requirement in each first environment image, and cuts the first partial image whose distortion degree is less than a set value, to obtain a plurality of first target images; The processor (2) compares the second reference network image with the standard image of the reference network, finds the second local image whose distortion degree meets the requirement in each second environment image, and cuts the second local image whose distortion degree is less than a set value, to obtain a plurality of second target images; The processor (2) splices the plurality of first target images and the plurality of second target images to obtain a target environment image for generating a three-dimensional image.
2. The vehicle front three-dimensional detection system based on a monocular camera as claimed in claim 1, characterized in that: Scaling the first environment images according to the sizes of the objects in the different first environment images so that the sizes of the objects in all the first environment images are consistent, and then acquiring the first target image; The second environment images are scaled according to the sizes of the objects in the different second environment images so that the sizes of the objects in all the second environment images are consistent, and then the second target image is acquired.
3. The vehicle front three-dimensional detection system based on a monocular camera as claimed in claim 2, characterized in that: It also includes a laser rangefinder (8), which is used to measure the distance from the first monocular camera and the second monocular camera to the front object.
4. The vehicle front three-dimensional detection system based on a monocular camera as claimed in claim 1, characterized in that: It also comprises a shell (9), wherein the left-right rotating platform (6) and the up-down rotating platform (7) are both installed in the shell (9), and the shell (9) has an open end, and the reference network is arranged in the open end.
5. The vehicle front three-dimensional detection system based on a monocular camera as claimed in claim 4, characterized in that: The left-right rotating platform (6) comprises a first mounting tube (61) and a support frame (62), the middle portion of the first mounting tube (61) is hinged to the support frame (62), one end of the first mounting tube (61) is connected to a driving mechanism (63), and the first monocular camera is mounted on the other end of the first mounting tube (61); The up-and-down rotating platform (7) includes a second mounting cylinder (71) and a push plate (72), the middle part of the second mounting cylinder (71) is hinged to the support frame (62), one end of the second mounting cylinder (71) is hinged to the push plate (72), and the second monocular camera is installed at the other end of the second mounting cylinder (71); the push plate (72) and the support frame (62) are slidably matched in the vertical direction, the lower surface of the push plate (72) is provided with an arc-shaped sliding groove, the bottom of the groove of the sliding groove is provided with a first inclined surface, the first mounting cylinder (61) is provided with a transmission block (64), the upper surface of the transmission block (64) is provided with a second inclined surface, the transmission block (64) is located in the sliding groove and the first inclined surface is in contact with the second inclined surface, and a compression spring (73) is provided between the upper surface of the push plate (72) and the top wall of the shell (9).
6. The vehicle front three-dimensional detection system based on a monocular camera as claimed in claim 5, characterized in that: The driving mechanism (63) comprises a linear motor, the linear motor is connected to a telescopic column (65), and the telescopic column (65) is hinged to the first mounting cylinder (61).
7. The vehicle front three-dimensional detection system based on a monocular camera as claimed in claim 4, 5 or 6, characterized in that: The open end of the housing (9) is provided with a transparent protective cover (10), and the reference network is arranged inside the protective cover (10).
8. The vehicle front three-dimensional detection system based on a monocular camera as claimed in claim 7, characterized in that: A positioning ring (11) is arranged on the outer side of the protective cover (10), a plurality of nozzles (12) are arranged on the positioning ring (11), the nozzles (12) are connected to a driving pump (13) via a pipeline, the driving pump (13) is connected to an air inlet pipe (14) and a water inlet pipe (15), valves are arranged on the air inlet pipe (14) and the water inlet pipe (15), and the water inlet pipe (15) is connected to a water tank (16).
9. The vehicle front three-dimensional detection system based on a monocular camera as claimed in claim 1, characterized in that: The first standard wire (5) and the second standard wire (4) are steel wires with a diameter of 0.1 mm, and a black coating is provided on the surface of the steel wires.
