A camera calibration verification method and system based on a video dark box

Abstract

The application provides a camera calibration and verification method and system based on a video dark box, which comprises the following steps: according to a preset front adjustment strategy, corresponding adjustment processing is performed on a first camera in the video dark box; according to a preset coarse calibration adjustment strategy, the yaw angle, the pitch angle and the roll angle of the first camera are respectively adjusted and processed through adjusting a clamp in the video dark box, and coarse calibration verification is performed; after determining that the coarse calibration verification is passed, according to a preset fine calibration adjustment strategy, the yaw angle, the pitch angle and the roll angle of the first camera are respectively adjusted and processed through adjusting the clamp, and fine calibration verification is performed; after determining that the fine calibration verification is passed, a new round of coarse calibration verification is performed. The problem that the relative position of the camera in the video dark box is not accurate enough, resulting in poor target recognition and functional performance effect is solved.

Description

TECHNICAL FIELD

[0001] The present application relates to the technical field of advanced auxiliary driving simulation testing, and particularly relates to a camera calibration and verification method and system based on a video dark box. BACKGROUND

[0002] An advanced driver assistance system (ADAS) is a general term of various systems that use sensors, communication devices, decision-making and execution devices, etc. installed on a vehicle to monitor a driver, the vehicle and its driving environment in real time, and assist the driver in performing driving tasks or actively avoid collision hazards through information and motion control. The existing ADAS hardware-in-the-loop (HIL) test scheme often uses a video dark box method for camera-in-the-loop testing. However, the camera needs to be calibrated before testing, and the accuracy of the calibration has a significant impact on target recognition and system performance.

[0003] The ordinary video dark box calibration method can be tested after passing the calibration. However, since the tolerance interval of successful camera calibration is large, and the camera calibration deviation angle and the camera measurement error are not directly related, the problem of "although the calibration is passed, the relative position of the camera in the dark box is not accurate enough, resulting in poor target recognition and performance" may occur, which may cause hidden troubles for the subsequent HIL function test. SUMMARY

[0004] The present application provides a camera calibration and verification method and system based on a video dark box, which is used to solve the problem of poor target recognition and performance caused by the inaccurate relative position of the camera in the dark box in the prior art.

[0005] In a first aspect, the present application provides a camera calibration and verification method based on a video dark box, comprising:

[0006] According to a preset front adjustment strategy, a first camera in the video dark box is subjected to corresponding adjustment processing;

[0007] According to a preset coarse calibration adjustment strategy, the yaw angle, pitch angle and roll angle of the first camera are subjected to coarse calibration adjustment processing by adjusting a clamp in the video dark box, and coarse calibration verification is performed;

[0008] After determining that the coarse calibration verification is passed, the yaw angle, pitch angle and roll angle of the first camera are subjected to fine calibration adjustment processing by adjusting the clamp according to a preset fine calibration adjustment strategy, and fine calibration verification is performed;

[0009] After determining that the fine calibration is passed, a new round of coarse calibration is performed.

[0010] In an embodiment, the adjusting the first camera in the video dark box according to the preset coarse calibration strategy includes:

[0011] A screen position is preset, and a field of view angle and an external parameter of a third camera in the scene simulation software are preset.

[0012] According to the field of view angle and the external parameter of the first camera in the video dark box, and according to the preset field of view angle and the external parameter of the third camera in the scene simulation software and information of the screen in the video dark box, a preset distance between the first camera and the screen is obtained, and the distance between the first camera and the screen is adjusted to the preset distance.

[0013] According to the parameters of the lens in the video dark box and information of the screen in the video dark box, the position of the lens is determined, and the degrees of freedom of the first camera and the jig in the video dark box in up-down, front-back and left-right directions are coarsely adjusted according to the position of the lens.

[0014] According to the camera image reading module, the target center target and the laser pen, the degrees of freedom of the first camera in up-down and left-right directions are finely adjusted, so that the center of the adjusted first camera coincides with the target center displayed on the screen in up-down and left-right directions.

[0015] In an embodiment, the adjusting the first camera in the video dark box according to the preset coarse calibration strategy includes:

[0016] The degrees of freedom of the first camera in the left-right direction are coarsely adjusted by adjusting the jig using the level tool.

[0017] The current first yaw angle calibration deviation angle, the first pitch angle calibration deviation angle and the first roll angle calibration deviation angle of the first camera are obtained by the camera calibration module.

[0018] If any of the first yaw angle calibration deviation angle, the first pitch angle calibration deviation angle and the first roll angle calibration deviation angle is not within the corresponding first threshold range, the first camera is adjusted by adjusting the jig, so that the second yaw angle calibration deviation angle, the second pitch angle calibration deviation angle and the second roll angle calibration deviation angle of the adjusted first camera are all within the corresponding first threshold range.

[0019] The second yaw angle, the second pitch angle and the second roll angle of the first camera are respectively calibrated statically by using the camera calibration module and the checkerboard target, and the first camera is adjusted by adjusting the jig according to the static calibration result obtained by the camera calibration module, so that the third yaw angle, the third pitch angle and the third roll angle of the first camera reacquired by the camera calibration module after adjustment are all within the corresponding second threshold range.

[0020] The third yaw angle, the third pitch angle and the third roll angle of the first camera are respectively calibrated dynamically by using the camera calibration module and the dynamic traffic scene projected on the screen in the video dark box, and the first camera is adjusted by adjusting the jig according to the dynamic calibration result obtained by the camera calibration module, so that the fourth yaw angle, the fourth pitch angle and the fourth roll angle of the first camera reacquired by the camera calibration module after adjustment are all within the corresponding third threshold range.

[0021] The coarse calibration verification includes:

[0022] The third yaw angle, the third pitch angle and the third roll angle of the first camera are respectively recalibrated statically by using the camera calibration module and the checkerboard target, and the static calibration result is obtained by using the camera calibration module.

[0023] It is judged whether the fifth yaw angle, the fifth pitch angle and the fifth roll angle obtained by using the camera calibration module are within the corresponding second threshold range, if not, the coarse adjustment is re-performed according to the preset coarse adjustment strategy; if yes, then:

[0024] The third yaw angle, the third pitch angle and the third roll angle of the first camera are respectively calibrated dynamically by using the camera calibration module and the dynamic traffic scene projected on the screen in the video dark box, and the dynamic calibration result is obtained by using the camera calibration module.

[0025] It is judged whether the sixth yaw angle, the sixth pitch angle and the sixth roll angle reacquired by using the camera calibration module are within the corresponding third threshold range, if not, the coarse adjustment is re-performed according to the preset coarse adjustment strategy; if yes, it is judged that the coarse calibration verification is passed, and the coarse calibration verification process is ended.

[0026] In a specific embodiment, the first camera yaw angle, pitch angle and roll angle are respectively adjusted by adjusting the jig according to the preset fine adjustment strategy, which includes:

[0027] When it is determined that the calibration deviation angle of the roll angle of the first camera after the coarse adjustment does not fall within the preset roll angle deviation threshold range, the roll angle is fine-tuned by the level gauge so that the calibration deviation angle of the roll angle obtained by the camera calibration module after the adjustment falls within the first preset roll angle deviation threshold range, and the deviation value of the roll angle measurement value from the horizontal reference value falls within the second preset roll angle deviation threshold range, at which time the roll degree of freedom is fixed;

[0028] The real vehicle on which the second camera is installed is set to travel at a preset speed on a real lane, and a real vehicle lane center keeping module on the real vehicle is turned on to ensure that the real vehicle always travels in the center of the lane. A real lane line fitting coefficient true value acquisition module is used to obtain the lane line fitting coefficient true value collected on the real lane, and a camera image reading module is used to obtain the lane line fitting coefficient measurement value detected by the second camera on the real lane.

[0029] The virtual vehicle on which the third camera is installed is set to travel at the preset speed on the virtual lane displayed on the screen in the video dark box, and an ideal lane center keeping module is turned on to ensure that the virtual vehicle always travels in the center of the lane. An ideal sensor module is used to obtain the lane line fitting coefficient true value collected on the virtual lane displayed on the screen in the video dark box, and the camera image reading module is used to obtain the lane line fitting coefficient measurement value detected by the first camera in the screen in the video dark box.

[0030] When the error between the lane line fitting coefficient true value collected by the real vehicle true value acquisition module and the lane line fitting coefficient true value collected by the ideal sensor module falls within the first preset fitting coefficient error threshold, and the error between the lane line fitting coefficient measurement value of the real lane detected by the second camera obtained by the camera image reading module and the lane line fitting coefficient measurement value of the virtual lane in the screen of the video dark box detected by the first camera obtained by the camera image reading module exceeds the second preset fitting coefficient error threshold range, the yaw angle of the first camera is adjusted so that the error between the two updated measurement values falls within the second preset fitting coefficient error threshold range.

[0031] The real vehicle on which the second camera is installed is set to be stationary in the center of the real lane, and the real object in front is also stationary in the center of the real lane. The real vehicle true value acquisition module is used to obtain the distance true value of the real vehicle from the real object in front, and the camera image reading module is used to obtain the distance measurement value of the real vehicle from the real object in front detected by the second camera.

[0032] The virtual host vehicle with the third camera installed is set to be static in the virtual lane center displayed on the screen in the video dark box, and the front virtual target is also static in the virtual host lane center displayed on the screen, the distance true value of the virtual host vehicle from the front virtual target collected on the dark box screen lane is obtained by using the ideal sensor module, and the distance measurement value of the virtual host vehicle from the front virtual target detected by the first camera in the video dark box is obtained by using the camera image reading module.

[0033] When the distance error between the distance true value of the virtual host vehicle from the front virtual target obtained by the ideal sensor module and the distance true value of the real host vehicle from the front real target obtained by the real vehicle true value collection module is within the first preset distance error threshold range, and the error between the distance measurement value of the real host vehicle from the front real target detected by the second camera and the distance measurement value of the virtual host vehicle from the front virtual target detected by the first camera is greater than the second preset distance error threshold range, the pitch angle of the first camera is adjusted so that the error between the updated distance measurement value of the real host vehicle from the front real target in the adjusted first camera and the distance measurement value of the virtual host vehicle from the front virtual target is within the second preset distance error threshold range.

[0034] The real host vehicle with the second camera installed is set to travel at a constant speed on the real lane center, and the front real target also travels at a constant speed on the real host lane center at a specific speed, the speed true value of the front real target is obtained by using the real vehicle true value collection module, and the speed measurement value of the front real target detected by the second camera is obtained by using the camera image reading module.

[0035] The virtual host vehicle with the third camera installed is set to travel at a constant speed on the virtual lane center displayed on the screen in the video dark box, and the front virtual target also travels at a constant speed on the virtual host lane center displayed on the screen at a specific speed, the speed true value of the front virtual target is obtained by using the ideal sensor module, and the speed measurement value of the front virtual target detected by the first camera in the video dark box is obtained by using the camera image reading module.

[0036] adjusting the virtual host vehicle speed and the virtual target speed in the scene simulation software to be consistent with the actual vehicle working condition, and making the error between the speed true value of the virtual target obtained by the ideal sensor module and the speed true value of the real target obtained by the actual vehicle true value acquisition module within a first preset speed error threshold range, and the error between the speed measurement value of the real target detected by the second camera and the speed measurement value of the virtual target in the video dark box screen detected by the first camera within a second preset speed error threshold range, adjusting the pitch angle of the first camera, so that the error between the updated speed measurement value of the real target and the speed measurement value of the virtual target is within the second preset speed error threshold range;

[0037] Then the fine calibration is performed, including:

[0038] determining whether the error between the lane line fitting coefficient true value collected by the actual vehicle true value acquisition module and the lane line fitting coefficient true value collected by the ideal sensor module is within a first preset fitting coefficient error threshold; if not, readjusting the position of the virtual host vehicle in the virtual lane in the virtual scene, and determining whether the error between the lane line fitting coefficient true value collected by the actual vehicle true value acquisition module and the lane line fitting coefficient true value collected by the ideal sensor module is within the first preset fitting coefficient error threshold; if yes, then:

[0039] determining whether the error between the lane line fitting coefficient measurement value of the real lane detected by the second camera and the lane line fitting coefficient measurement value of the virtual lane detected by the first camera is within a second preset fitting coefficient error threshold range; if not, readjusting the clamp according to the preset fine calibration strategy; if yes, then:

[0040] determining whether the error between the distance true value of the real host vehicle from the front real target obtained by the actual vehicle true value acquisition module and the distance true value of the virtual host vehicle from the front virtual target obtained by the ideal sensor module is within a first preset distance error threshold range; if not, readjusting the position of the host vehicle and the front vehicle in the virtual scene, and determining whether the error between the distance true value of the real host vehicle from the front real target obtained by the actual vehicle true value acquisition module and the distance true value of the virtual host vehicle from the front virtual target obtained by the ideal sensor module is within the first preset distance error threshold range; if yes, then:

[0041] determining whether an error between a distance measurement value of a real vehicle from a real object in front acquired by the second camera and a distance measurement value of a virtual vehicle from a virtual object in front acquired by the first camera is within a second preset distance error threshold range; if not, adjusting the fixture according to a preset fine adjustment strategy; if yes, determining that the fine adjustment verification passes, and ending the fine adjustment verification process.

[0042] determining whether an error between a speed true value of the real object in front acquired by the real vehicle true value acquisition module and a speed true value of the virtual object in front acquired by the ideal sensor module is within a first preset speed error threshold range; if not, adjusting the speed of the vehicle and the front vehicle in the virtual scene, and determining whether an error between a speed true value of the real object in front acquired by the real vehicle true value acquisition module and a speed true value of the virtual object in front acquired by the ideal sensor module is within the first preset speed error threshold range; if yes, adjusting the speed of the vehicle and the front vehicle in the virtual scene, and determining whether an error between a speed true value of the real object in front acquired by the real vehicle true value acquisition module and a speed true value of the virtual object in front acquired by the ideal sensor module is within the first preset speed error threshold range; if yes, determining that the fine adjustment verification passes, and ending the fine adjustment verification process.

[0043] determining whether an error between a speed measurement value of the real object in front acquired by the second camera and a speed measurement value of the virtual object in front acquired by the first camera is within a second preset speed error threshold range; if not, adjusting the fixture according to a preset fine adjustment strategy; if yes, determining that the fine adjustment verification passes, and ending the fine adjustment verification process.

[0044] In a second aspect, the application provides a camera calibration verification system based on a video dark box, comprising: a camera image reading module, a real vehicle true value acquisition module, an ideal sensor module, a camera calibration module, an execution module, a real vehicle lane centering and keeping module, an ideal lane centering and keeping module, a checkerboard target and a target center target.

[0045] The camera calibration module is configured to store a preset coarse adjustment strategy, perform static calibration and dynamic calibration, and acquire static calibration results and dynamic calibration results, and acquire a yaw angle calibration deviation angle, a pitch angle calibration deviation angle and a roll angle calibration deviation angle after calibration.

[0046] The execution module is configured to store a pre-adjustment strategy and a fine adjustment strategy, perform corresponding adjustment processing on the first camera in the video dark box according to the preset pre-adjustment strategy, perform coarse adjustment processing on the yaw angle, the pitch angle and the roll angle of the first camera by adjusting the fixture in the video dark box according to the preset coarse adjustment strategy, and perform coarse adjustment verification, perform fine adjustment processing on the yaw angle, the pitch angle and the roll angle of the first camera by adjusting the fixture according to the preset fine adjustment strategy after determining that the coarse adjustment verification passes, and perform fine adjustment verification, and perform a new round of coarse adjustment verification after determining that the fine adjustment verification passes.

[0047] In an embodiment, the execution module is specifically configured to:

[0048] According to the field of view angle and the extrinsic parameter of the first camera in the video dark box, and according to the field of view angle and the extrinsic parameter of the third camera preset in the scene simulation software and the information of the screen in the video dark box, a preset distance between the first camera and the screen is obtained, and the distance between the first camera and the screen is adjusted to the preset distance.

[0049] According to the parameters of the lens in the video dark box and the information of the screen in the video dark box, the position of the lens is determined, and the degrees of freedom of the first camera and the clamp in the video dark box in the up-down, front-back and left-right directions are coarsely adjusted according to the position of the lens.

[0050] According to the camera image reading module, the bullseye target and the laser pen, the degrees of freedom of the first camera in the up-down and left-right directions are finely adjusted so that the center of the adjusted first camera coincides with the bullseye displayed on the screen in the up-down and left-right directions.

[0051] The bullseye target is used to adjust the up-down and left-right positions of the first camera and the clamp to be on the same horizontal line and the same vertical line as the bullseye displayed on the screen.

[0052] In an embodiment, the execution module is specifically configured to:

[0053] The horizontal tool is used to coarsely adjust the roll degree of freedom of the first camera by adjusting the clamp.

[0054] The camera calibration module is specifically configured to obtain the first yaw angle calibration deviation angle, the first pitch angle calibration deviation angle and the first roll angle calibration deviation angle of the first camera.

[0055] If the camera calibration module determines that any of the first yaw angle calibration deviation angle, the first pitch angle calibration deviation angle and the first roll angle calibration deviation angle is not within the corresponding first threshold range, the execution module is specifically configured to adjust the first camera by adjusting the clamp, so that the second yaw angle calibration deviation angle, the second pitch angle calibration deviation angle and the second roll angle calibration deviation angle of the adjusted first camera are all within the corresponding first threshold range.

[0056] The camera calibration module calibrates the second yaw angle, the second pitch angle and the second roll angle of the first camera respectively by using the checkerboard target; and the execution module is specifically configured to: according to the static calibration result obtained by the camera calibration module, adjust the first camera by adjusting the clamp, so that the third yaw angle, the third pitch angle and the third roll angle of the first camera reacquired by the camera calibration module after the adjustment are all within the corresponding second threshold range.

[0057] The camera calibration module calibrates the second yaw angle, the second pitch angle and the second roll angle of the first camera respectively by using the checkerboard target; and the execution module is specifically configured to: according to the static calibration result obtained by the camera calibration module, adjust the first camera by adjusting the clamp, so that the third yaw angle, the third pitch angle and the third roll angle of the first camera reacquired by the camera calibration module after the adjustment are all within the corresponding second threshold range.

[0058] The checkerboard target is used for static calibration in coarse calibration adjustment and static calibration in coarse calibration verification.

[0059] In a specific embodiment, the execution module is specifically configured to: when the camera calibration module determines that the roll angle of the first camera after the coarse calibration adjustment is not within the preset roll angle deviation threshold range, finely adjust the roll angle by using a level, so that the roll angle deviation angle of the roll angle obtained by the camera calibration module after the adjustment is within the first preset roll angle deviation threshold range, and the deviation value of the roll angle measurement value from the horizontal reference value is within the second preset roll angle deviation threshold range.

[0060] The real vehicle true value acquisition module is configured to acquire lane line fitting coefficient true values on a real lane when a real ego vehicle carrying a second camera drives on a real lane center;

[0061] The ideal sensor module is configured to acquire lane line fitting coefficient true values on a virtual lane detected by a third camera when a virtual ego vehicle carrying the third camera drives on a virtual lane center;

[0062] The camera image reading module is configured to acquire lane line fitting coefficient measurement values on a real lane detected by a second camera when a real ego vehicle carrying the second camera drives on a real lane center;

[0063] The camera image reading module is further configured to acquire virtual lane line fitting coefficient measurement values on a dark box screen detected by the first camera in the video dark box when the virtual ego vehicle carrying the third camera drives on the virtual lane center;

[0064] The real vehicle lane centering keeping module is configured to keep the real ego vehicle equipped with the second camera always driving in the center of the real lane.

[0065] The ideal lane centering keeping module is configured to keep the virtual ego vehicle equipped with the third camera always driving in the center of the virtual lane.

[0066] The execution module is further configured to adjust the yaw angle of the first camera when the error between the lane line fitting coefficient true value collected by the real vehicle true value collection module and the lane line fitting coefficient true value collected by the ideal sensor module is within the first preset fitting coefficient error threshold, and when the error between the lane line fitting coefficient measurement value of the real lane detected by the second camera and the lane line fitting coefficient measurement value of the virtual lane in the video dark box screen detected by the first camera exceeds the second preset fitting coefficient error threshold range, so that the error between the two updated measurement values is within the second preset fitting coefficient error threshold range.

[0067] The real vehicle true value collection module is further configured to obtain the distance true value of the real ego vehicle from the front real object when the real ego vehicle equipped with the second camera is stationary in the center of the real lane, and the front real object is also stationary in the center of the real lane.

[0068] The ideal sensor module is further configured to obtain the distance true value of the virtual ego vehicle from the front virtual object detected by the third camera when the virtual ego vehicle equipped with the third camera is stationary in the center of the virtual lane displayed on the video dark box screen, and the front virtual object is also stationary in the center of the virtual lane.

[0069] The camera image reading module is configured to obtain the distance measurement value of the real ego vehicle from the front real object detected by the second camera when the real ego vehicle equipped with the second camera is stationary in the center of the real lane, and the front real object is also stationary in the center of the real lane.

[0070] The camera image reading module is further configured to obtain the distance measurement value of the virtual ego vehicle from the front virtual object on the video dark box screen detected by the first camera when the virtual ego vehicle equipped with the third camera is stationary in the center of the virtual lane displayed on the video dark box screen, and the front virtual object is also stationary in the center of the virtual lane.

[0071] The execution module is specifically configured to, when the error between the distance true value of the real object target in front collected by the real car true value collection module and the distance true value of the virtual object target in front collected by the ideal sensor module is within a first preset distance error threshold, and the error between the distance measurement value of the real car in front detected by the second camera and the distance measurement value of the virtual car in front detected by the first camera in the video dark box screen obtained by the camera image reading module exceeds a second preset distance error threshold range, adjust the pitch angle of the first camera, so that the error between the updated distance measurement value of the real car in front in the adjusted first camera and the distance measurement value of the virtual car in front is within the second preset distance error threshold range.

[0072] The real car true value collection module is further configured to, when the real car carrying the second camera uniformly travels at a set speed on the real lane center, and the real object target in front also uniformly travels at a specific speed on the real lane center, obtain the speed true value of the real object target in front on the real road detected by the second camera.

[0073] The ideal sensor module is further configured to, when the virtual car carrying the third camera uniformly travels at a set speed on the virtual lane center displayed on the video dark box screen, and the virtual object target in front also uniformly travels at a specific speed on the virtual lane center, obtain the speed true value of the virtual object target in front detected by the third camera.

[0074] The camera image reading module is configured to, when the real car carrying the second camera uniformly travels at a set speed on the real lane center, and the real object target in front also uniformly travels at a specific speed on the real lane center, obtain the speed measurement value of the real object target in front on the real lane detected by the second camera.

[0075] The camera image reading module is further configured to, when the virtual car carrying the third camera uniformly travels at a set speed on the virtual lane center displayed on the video dark box screen, and the virtual object target in front also uniformly travels at a specific speed on the virtual lane center, obtain the speed measurement value of the virtual object target in front on the dark box screen detected by the first camera.

[0076] The execution module is specifically configured to adjust the pitch angle of the first camera so that the error between the updated speed measurement value of the front real object and the speed measurement value of the front virtual object is within the second preset speed error threshold range when the error between the speed true value of the front real object collected by the real vehicle true value collection module and the speed true value of the front virtual object collected by the ideal sensor module is within the first preset speed error threshold, and the error between the speed measurement value of the front real object detected by the second camera and the speed measurement value of the front virtual object in the video dark box screen detected by the first camera exceeds the second preset speed error threshold range.

[0077] The application provides a camera calibration and verification method and system based on a video dark box, which comprises the following steps: according to a preset pre-adjustment strategy, a first camera in the video dark box is subjected to corresponding adjustment processing; according to a preset coarse calibration adjustment strategy, a clamp in the video dark box is adjusted to perform coarse calibration adjustment processing on the yaw angle, the pitch angle and the roll angle of the first camera respectively, and coarse calibration verification is performed; after it is determined that the coarse calibration verification is passed, according to a preset fine calibration adjustment strategy, the clamp is adjusted to perform fine calibration adjustment processing on the yaw angle, the pitch angle and the roll angle of the first camera respectively, and fine calibration verification is performed; after it is determined that the fine calibration verification is passed, a new round of coarse calibration verification is performed. Compared with the prior art, although calibration is passed, the relative position of the camera in the dark box is not accurate enough, resulting in poor target recognition and functional performance. After pre-adjustment and coarse calibration adjustment, coarse calibration verification is performed, and according to the preset fine calibration adjustment strategy, fine calibration adjustment and corresponding verification processing are performed on the yaw angle, the pitch angle and the roll angle of the camera in the video dark box, so that the accuracy of the relative position of the camera in the dark box is effectively improved, and the target recognition and functional performance are improved. BRIEF DESCRIPTION OF DRAWINGS

[0078] In order to more clearly illustrate the technical solutions in the embodiments of the application or the prior art, the following will briefly introduce the drawings needed to be used in the embodiments or the prior art description. Obviously, the drawings in the following description are some embodiments of the application, and other drawings can be obtained by those skilled in the art without creative labor.

[0079] Figure 1 A flowchart of a camera calibration and verification method based on a video dark box according to an embodiment of the application is provided.

[0080] Figure 2 A flowchart of a camera calibration and verification method based on a video dark box according to an embodiment of the application is provided.

[0081] Figure 3a Figure 3 is a flowchart illustrating an embodiment of a camera calibration and verification method based on a video dark box according to the present application;

[0082] Figure 3b Figure 4 is a flowchart illustrating an embodiment of a camera calibration and verification method based on a video dark box according to the present application;

[0083] Figure 4a and Figure 4b Figure 5 is a flowchart illustrating an embodiment of a camera calibration and verification method based on a video dark box according to the present application;

[0084] Figure 4c Figure 6 is a flowchart illustrating an embodiment of a camera calibration and verification method based on a video dark box according to the present application;

[0085] Figure 5 Figure 7 is a structural diagram illustrating an embodiment of a camera calibration and verification system according to the present application. DETAILED DESCRIPTION

[0086] In order to make the objectives, technical solutions and advantages of the embodiments of the present application clearer, the technical solutions in the embodiments of the present application will be described below in connection with the drawings in the embodiments of the present application. Obviously, the described embodiments are only some of the embodiments of the present application, rather than all the embodiments of the present application. Based on the embodiments in the present application, all other embodiments made by those of ordinary skill in the art under the inspiration of the present embodiments are within the scope of protection of the present application.

[0087] The terms "first", "second", "third", "fourth", "fifth", "sixth" and the like (if any) in the description, claims and drawings of the present application are used to distinguish similar objects, and do not necessarily indicate a specific order or sequence. It should be understood that the data thus used can be interchanged under appropriate circumstances, so that the embodiments of the present application described herein can be implemented in an order other than that illustrated or described herein. In addition, the terms "include" and "have" and any variations thereof are intended to cover non-exclusive inclusion, for example, a process, method, system, product or device that includes a series of steps or units does not necessarily have to be limited to those steps or units clearly listed, but can include other steps or units that are not clearly listed or inherent to these processes, methods, products or devices.

[0088] The solutions of the present application will be described in detail below.

[0089] The camera calibration and verification method based on the video dark box specifically comprises: according to a preset pre-adjustment strategy, performing corresponding adjustment processing on the first camera in the video dark box; according to a preset coarse calibration adjustment strategy, performing coarse calibration adjustment processing on the yaw angle, the pitch angle and the roll angle of the first camera by adjusting the clamp in the video dark box, and performing coarse calibration verification; after determining that the coarse calibration verification is passed, performing fine calibration adjustment processing on the yaw angle, the pitch angle and the roll angle of the first camera by adjusting the clamp according to a preset fine calibration adjustment strategy, and performing fine calibration verification; after determining that the fine calibration verification is passed, performing a new round of coarse calibration verification.

[0090] In the embodiment, the first camera is a camera in the video dark box, which is used to acquire an image displayed on a screen in the video dark box. The clamp is installed on a camera support in the video dark box, and is used to adjust the yaw angle, the pitch angle and the roll angle of the first camera.

[0091] A specific implementation mode of the camera calibration and verification method based on the video dark box is shown in Figure 1 . Figure 1 A flowchart of a first embodiment of the camera calibration and verification method based on the video dark box is provided in the present application. Referring to Figure 1 , the camera calibration and verification method based on the video dark box specifically comprises the following steps:

[0092] Step S101: a pre-adjustment process is performed, that is, according to a preset pre-adjustment strategy, corresponding adjustment processing is performed on the first camera in the video dark box.

[0093] Step S102: a coarse calibration adjustment process is performed, that is, according to a preset coarse calibration adjustment strategy, coarse calibration adjustment processing is performed on the yaw angle, the pitch angle and the roll angle of the first camera by adjusting the clamp in the video dark box.

[0094] Step S103: a coarse calibration verification process is performed. If the coarse calibration verification process is not passed, step S102 is executed, that is, the coarse calibration adjustment process is performed again; if the coarse calibration verification process is passed, step S104 is executed.

[0095] Step S104: a fine calibration adjustment process is entered, that is, according to a preset fine calibration adjustment strategy, fine calibration adjustment processing is performed on the yaw angle, the pitch angle and the roll angle of the first camera by adjusting the clamp.

[0096] Step S105: after the fine calibration adjustment process, a fine calibration verification process is performed. If the fine calibration verification process is not passed, step S104 is executed, that is, the fine calibration adjustment process is performed again; if the fine calibration verification process is passed, step S106 is executed, that is, a new round of coarse calibration verification process is performed.

[0097] Step S106: Perform a new round of coarse calibration verification. Determine whether the coarse calibration verification process passes. If it passes, end the overall calibration verification process. If it fails, execute step S102, that is, repeat the coarse calibration adjustment process.

[0098] In this embodiment, the first camera in the video darkroom is adjusted according to a preset pre-adjustment strategy. According to a preset coarse calibration strategy, the yaw, pitch, and tilt angles of the first camera are coarsely adjusted by adjusting the fixtures in the video darkroom, and then coarse calibration is verified. After the coarse calibration is confirmed to be passed, the yaw, pitch, and tilt angles of the first camera are finely adjusted by adjusting the fixtures according to a preset fine calibration strategy, and then fine calibration is verified. After the fine calibration is confirmed to be passed, a new round of coarse calibration is performed. Compared to existing technologies, although the calibration is successful, the target recognition and functional performance are poor due to the inaccurate relative position of the camera in the dark box. This application performs coarse calibration verification after pre-adjustment and coarse calibration adjustment, and performs fine calibration adjustment and corresponding verification processing on the yaw angle, pitch angle and side tilt angle of the camera in the video dark box according to the preset fine calibration adjustment strategy. This effectively improves the accuracy of the relative position of the camera in the dark box, thereby improving the target recognition and functional performance.

[0099] Figure 2 This is a flowchart illustrating a second embodiment of a camera calibration and verification method based on a video darkroom provided in this application. Figure 1 Based on the illustrated embodiment, see also Figure 2 The above step S101 specifically includes the following steps:

[0100] Step S201: Preset the screen position and preset the field of view and external parameters of the third camera in the scene simulation software.

[0101] Step S202: Based on the field of view and extrinsic parameters of the first camera in the video dark box, and based on the field of view and extrinsic parameters of the third camera preset in the scene simulation software, as well as the information of the screen in the video dark box, obtain the preset distance between the first camera and the screen, and adjust the distance between the first camera and the screen to the preset distance.

[0102] Step S203: Determine the position of the lens based on the parameters of the lens in the video dark box and the information of the screen in the video dark box, and make coarse adjustments to the degrees of freedom of the first camera and the clamp in the video dark box in the three directions of up and down, front and back, and left and right based on the position of the lens.

[0103] Step S204: Based on the camera image reading module, the bullseye target, and the laser pointer, finely adjust the vertical and horizontal degrees of freedom of the first camera so that the center of the adjusted first camera coincides with the bullseye displayed on the screen in both vertical and horizontal directions.

[0104] In this embodiment, according to the preset pre-adjustment strategy, the first camera in the video dark box is adjusted accordingly, that is, the pre-process is executed: (1) First adjust the screen position and configure the camera field of view (FOV) and external parameters in the scene simulation software; (2) Calculate the distance between the camera and the screen according to the camera's internal and external parameters, screen length and width, and resolution; (3) Determine the lens position according to the lens parameters and fine-tune the camera and fixture positions; (4) Use the camera image reading module, calibration target and laser pen to adjust the horizontal and vertical positions of the camera to ensure that the camera and the target displayed on the screen are on the same horizontal line and the same vertical line.

[0105] In this embodiment, the first camera in the video darkroom is adjusted according to a preset pre-adjustment strategy, providing the prerequisite for subsequent coarse and fine calibration adjustments.

[0106] In the above Figures 1 to 2 Based on the embodiment shown, step S102 is the coarse standard adjustment process, and step S103 is the coarse standard verification process.

[0107] Specifically, the coarse adjustment process includes: (1) using a level tool to adjust the clamps to achieve coarse adjustment of the tilt freedom of the first camera.

[0108] (2) Use the camera calibration module to obtain the first yaw angle calibration deviation angle, the first pitch angle calibration deviation angle and the first tilt angle calibration deviation angle of the first camera.

[0109] (3) If any of the calibration deviation angles of the first yaw angle, the first pitch angle, and the first tilt angle are not within the corresponding first threshold range, the first camera is adjusted by adjusting the fixture so that the second yaw angle, the second pitch angle, and the second tilt angle of the adjusted first camera are all within the corresponding first threshold range.

[0110] (4) The second yaw angle, the second pitch angle and the second roll angle of the first camera are respectively statically calibrated by using the camera calibration module and the checkerboard target, and according to the static calibration result obtained by the camera calibration module, the first camera is adjusted by adjusting the clamp, so that the third yaw angle, the third pitch angle and the third roll angle of the first camera reacquired by using the camera calibration module after adjustment are all within the corresponding second threshold range.

[0111] (5) The third yaw angle, the third pitch angle and the third roll angle of the first camera are respectively dynamically calibrated by using the camera calibration module and the dynamic traffic scene projected on the screen in the video dark box, and according to the dynamic calibration result obtained by the camera calibration module, the first camera is adjusted by adjusting the clamp, so that the fourth yaw angle, the fourth pitch angle and the fourth roll angle of the first camera reacquired by using the camera calibration module after adjustment are all within the corresponding third threshold range.

[0112] A specific implementation of the coarse calibration adjustment process is shown in Figure 3a . Figure 3a An embodiment three of a camera calibration and verification method based on a video dark box provided in the present application is shown in Figure 3a . In the present embodiment, according to a preset coarse calibration adjustment strategy, the yaw angle, the pitch angle and the roll angle of the first camera are respectively processed by using the clamp in the video dark box, which specifically includes the following steps:

[0113] Step S301: The level of the clamp is adjusted by using the level, and the coarse adjustment of the roll degree of freedom of the first camera is realized by adjusting the clamp.

[0114] Step S302: The camera calibration module is used to adjust the clamp, so that the calibration deviation values ξ 1a , ξ 1b , ξ 1c of the yaw angle, the pitch angle and the roll angle of the camera are respectively within the threshold ranges δ 1a , δ 1b , δ 1c .

[0115] Step S303: The camera calibration module and the checkerboard target are used to perform static calibration, and according to the static calibration result, the clamp is fine-adjusted, so that the calibration deviation angles ξ 2a , ξ 2b , ξ 2c of the yaw angle, the pitch angle and the roll angle are respectively within the threshold ranges δ 2a , δ 2b , δ 2c .

[0116] Step S304: dynamic calibration is performed by using the camera calibration module and the dynamic traffic scene projected on the screen in the video dark box, and the fixture in the video dark box is adjusted according to the calibration result, so that the calibration deviation angles ξ 3a 、ξ 3b 、ξ 3c of the yaw angle, the pitch angle and the roll angle are respectively within the threshold value δ 3a 、δ 3b 、δ 3c .

[0117] In this embodiment, according to the preset coarse calibration adjustment strategy, the yaw angle, the pitch angle and the roll angle of the first camera are respectively coarsely calibrated and adjusted by adjusting the fixture in the video dark box by using the level, the camera calibration module, the checkerboard target and the dynamic traffic scene projected on the screen in the video dark box, which provides a prerequisite for subsequent fine calibration adjustment.

[0118] Specifically, the coarse calibration verification process specifically includes:

[0119] (1) The third yaw angle, the third pitch angle and the third roll angle of the first camera are respectively statically calibrated by using the camera calibration module and the checkerboard target, and the static calibration result is obtained by using the camera calibration module.

[0120] (2) It is judged whether the fifth yaw angle calibration deviation angle, the fifth pitch angle calibration deviation angle and the fifth roll angle calibration deviation angle obtained by using the camera calibration module are within the corresponding second threshold value range, if not, the coarse calibration adjustment is performed again according to the preset coarse calibration adjustment strategy, if yes, the third yaw angle, the third pitch angle and the third roll angle of the first camera are respectively dynamically calibrated by using the camera calibration module and the dynamic traffic scene projected on the screen in the video dark box, and the dynamic calibration result is obtained by using the camera calibration module.

[0121] (3) It is judged whether the sixth yaw angle calibration deviation angle, the sixth pitch angle calibration deviation angle and the sixth roll angle calibration deviation angle reobtained by using the camera calibration module are within the corresponding third threshold value range, if not, the coarse calibration adjustment is performed again according to the preset coarse calibration adjustment strategy, if yes, it is determined that the coarse calibration verification is passed, and the coarse calibration verification process is ended.

[0122] A specific implementation of the coarse calibration verification process is shown in Figure 3b . Figure 3b A flowchart of an embodiment four of the camera calibration verification method based on the video dark box provided by the present application is shown in Figure 3b , and the coarse calibration verification process is executed as follows:

[0123] Step S361: Static calibration is performed by using the camera calibration module and the checkerboard target, and the static calibration result is obtained.

[0124] Step S362: Determine the calibration deviation angle ξ for yaw, pitch, and roll angles. 4a ξ 4b ξ 4c Is it at the threshold δ? 2a δ 2b δ 2c Within the range. If yes, proceed to step S363; otherwise, return to the coarse standard adjustment process.

[0125] Step S363: Perform dynamic calibration using the camera calibration module and the dynamic traffic scene projected on the screen in the video dark box, and obtain the dynamic calibration results.

[0126] Step S364: Determine the calibration deviation angle ξ for yaw, pitch, and roll angles. 5a ξ 5b ξ 5c Is it at the threshold δ? 3a δ 3b δ 3c If the range is within the specified range, the coarse label verification process ends; otherwise, the process returns to the coarse label adjustment process.

[0127] In the above Figures 1 to 3b Based on the illustrated embodiment, step S104 above is the fine-calibration adjustment process, which specifically includes the following steps:

[0128] (1) When the calibration deviation angle of the tilt angle of the first camera after coarse calibration is determined to be outside the preset tilt angle deviation threshold, the tilt angle is finely adjusted by the level so that the calibration deviation angle of the tilt angle obtained by the camera calibration module after adjustment is within the first preset tilt angle deviation threshold, and the deviation between the tilt angle measurement value and the horizontal reference value is within the second preset tilt angle deviation threshold. At this time, the tilt degree of freedom is fixed.

[0129] (2) Set the real vehicle with the second camera installed to drive at a preset speed in the real lane, and turn on the real vehicle lane centering module to ensure that the real vehicle always drives in the center of the lane; use the real vehicle true value acquisition module to obtain the true value of the lane line fitting coefficient collected in the real lane, and use the camera image reading module to obtain the measured value of the lane line fitting coefficient detected by the second camera in the real lane.

[0130] (3) setting the virtual host vehicle with the third camera installed to drive at a preset speed on the virtual lane displayed on the screen in the video dark box at a constant speed, starting the ideal lane center keeping module to ensure that the virtual host vehicle always drives in the center of the lane; using the ideal sensor module to obtain the lane line fitting coefficient true value collected on the virtual lane displayed on the screen in the video dark box, and using the camera image reading module to obtain the lane line fitting coefficient measurement value detected by the first camera in the video dark box.

[0131] (4) when the error between the lane line fitting coefficient true value collected by the real vehicle true value collection module and the lane line fitting coefficient true value collected by the ideal sensor module is within the first preset fitting coefficient error threshold, and the error between the lane line fitting coefficient measurement value of the real lane detected by the second camera obtained by the camera image reading module and the lane line fitting coefficient measurement value of the virtual lane in the screen of the video dark box detected by the first camera obtained by the camera image reading module exceeds the second preset fitting coefficient error threshold range, adjusting the yaw angle of the first camera so that the error between the two updated measurement values is within the second preset fitting coefficient error threshold range.

[0132] (5) setting the real host vehicle with the second camera installed to be stationary in the center of the real lane, and the front real object is also stationary in the center of the real host lane, using the real vehicle true value collection module to obtain the distance true value of the real host vehicle from the front real object, and using the camera image reading module to obtain the distance measurement value of the real host vehicle from the front real object detected by the second camera.

[0133] (6) setting the virtual host vehicle with the third camera installed to be stationary in the center of the virtual lane displayed on the screen in the video dark box, and the front virtual target is also stationary in the center of the virtual host lane displayed on the screen, using the ideal sensor module to obtain the distance true value of the virtual host vehicle from the front virtual target collected on the dark box screen lane, and using the camera image reading module to obtain the distance measurement value of the virtual host vehicle from the front virtual target detected by the first camera in the video dark box.

[0134] (7) adjusting the virtual ego vehicle position in the scene simulation software so that the distance error between the distance true value of the virtual ego vehicle to the virtual target in front obtained by the ideal sensor module and the distance true value of the real ego vehicle to the real target in front obtained by the real vehicle true value acquisition module is within a first preset distance error threshold range, and the error between the distance measurement value of the real ego vehicle to the real target in front detected by the second camera and the distance measurement value of the virtual ego vehicle to the virtual target in front in the video dark box screen detected by the first camera exceeds a second preset distance error threshold range, adjusting the pitch angle of the first camera so that the error between the updated distance measurement value of the real ego vehicle to the real target in front in the adjusted first camera and the distance measurement value of the virtual ego vehicle to the virtual target in front is within the second preset distance error threshold range.

[0135] (8) setting the real ego vehicle on which the second camera is installed to uniformly move at a set speed on the real lane center, and the real target in front also uniformly moves at a specific speed on the real lane center, obtaining the speed true value of the real target in front by the real vehicle true value acquisition module, and obtaining the speed measurement value of the real target in front detected by the second camera by the camera image reading module.

[0136] (9) setting the virtual ego vehicle on which the third camera is installed to uniformly move at a set speed on the virtual lane center displayed on the screen in the video dark box, and the virtual target in front also uniformly moves at a specific speed on the virtual lane center displayed on the screen, obtaining the speed true value of the virtual target in front by the ideal sensor module, and obtaining the speed measurement value of the virtual target in front detected by the first camera in the video dark box by the camera image reading module.

[0137] (10) adjusting the virtual ego vehicle speed and the virtual target in front speed in the scene simulation software to be consistent with the real vehicle working condition respectively, and when the error between the speed true value of the virtual target in front obtained by the ideal sensor module and the speed true value of the real target in front obtained by the real vehicle true value acquisition module is within a first preset speed error threshold range, and the error between the speed measurement value of the real target in front detected by the second camera and the speed measurement value of the virtual target in front in the video dark box screen detected by the first camera exceeds a second preset speed error threshold range, adjusting the pitch angle of the first camera so that the error between the updated speed measurement value of the real target in front and the speed measurement value of the virtual target in front is within the second preset speed error threshold range.

[0138] In the embodiment, the second camera is arranged at the front windshield of the vehicle.

[0139] The real self-lane refers to the same real lane as the real experimental vehicle with the second camera installed; and the virtual self-lane refers to the same virtual lane as the virtual experimental vehicle with the third camera installed.

[0140] When the real vehicle true value collection experiment is performed, the distances can be set as 10 m, 20 m, 30 m, 40 m, 50 m, 60 m, 70 m, 80 m, 90 m, 100 m, 110 m, 120 m, 130 m, 140 m and 150 m in sequence.

[0141] One specific implementation of the fine calibration adjustment process is shown in Figure 4a and Figure 4b . Figure 4a and Figure 4b is a flowchart of one embodiment of the camera calibration verification method based on the video dark box provided in the present application, which is shown in Figure 4a and Figure 4b . The fine calibration adjustment process specifically includes the following steps:

[0142] Step S401: The level is used to finely adjust the roll angle, so that the calibration deviation angle of the roll angle obtained by using the camera calibration module after adjustment is within the first preset roll angle deviation threshold range, and the deviation value of the roll angle measurement value from the horizontal reference value is within the second preset roll angle deviation threshold range, at which time the roll degree of freedom is fixed.

[0143] Step S402: The real self-vehicle with the second camera installed is set to travel at a preset speed on the real lane, and the real vehicle lane center keeping module on the real self-vehicle is turned on to ensure that the real self-vehicle always travels in the center of the lane; the lane line fitting coefficient true value collected on the real lane is obtained by using the real vehicle true value collection module, and the lane line fitting coefficient measurement value detected by the second camera on the real lane is obtained by using the camera image reading module.

[0144] Step S403: The virtual self-vehicle with the third camera installed is set to travel at a preset speed on the virtual lane displayed on the screen in the video dark box, and the ideal lane center keeping module is turned on to ensure that the virtual self-vehicle always travels in the center of the lane; the lane line fitting coefficient true value collected on the virtual lane displayed on the screen in the video dark box is obtained by using the ideal sensor module, and the lane line fitting coefficient measurement value detected by the first camera in the video dark box on the screen is obtained by using the camera image reading module.

[0145] Step S404: It is judged whether the error between the lane line fitting coefficient true value collected by using the real vehicle true value collection module and the lane line fitting coefficient true value collected by using the ideal sensor module is within the first preset fitting coefficient error threshold; if not, return to step S403; if yes, execute step S405.

[0146] Step S405: determining whether the error between the lane line fitting coefficient measurement value of the real lane line detected by the second camera and the lane line fitting coefficient measurement value of the virtual lane line in the video dark box screen detected by the first camera is within a second preset fitting coefficient error threshold range; if not, adjusting the yaw angle of the first camera and returning to step S405; if yes, executing step S406.

[0147] Step S406: setting the real ego vehicle equipped with the second camera to be stationary in the real lane center and the front real object to be stationary in the real lane center, acquiring the distance true value A1 of the real ego vehicle from the front real object by the real vehicle true value acquisition module, and acquiring the distance measurement value a1 of the real ego vehicle from the front real object detected by the second camera by the camera image reading module.

[0148] Step S407: setting the virtual ego vehicle equipped with the third camera to be stationary in the virtual lane center and the front virtual object to be stationary in the virtual lane center, acquiring the distance true value B1 of the virtual ego vehicle from the front virtual object by the ideal sensor module, and acquiring the distance measurement value b1 of the virtual ego vehicle from the front virtual object detected by the first camera in the video dark box by the camera image reading module.

[0149] Step S408: determining whether the error between the distance true value A1 and the distance true value B1 is within a first preset distance error threshold range; if not, returning to step S407; if yes, executing step S409.

[0150] Step S409: determining whether the error between the distance measurement value a1 and the distance measurement value b1 is within a second preset distance error threshold range; if not, adjusting the pitch angle of the first camera and returning to step S409; if yes, executing step S410.

[0151] Step S410: setting the real ego vehicle equipped with the second camera to travel at a set speed on the real lane center and the front real object to travel at a specific speed on the real lane center, acquiring the speed true value C1 of the front real object by the real vehicle true value acquisition module, and acquiring the speed measurement value c1 of the front real object detected by the second camera by the camera image reading module.

[0152] Step S411: setting the virtual ego vehicle equipped with the third camera to travel at a set speed on the screen displayed lane center in the video dark box and the front virtual object to travel at a specific speed on the screen displayed lane center, acquiring the speed true value D1 of the front virtual object by the ideal sensor module, and acquiring the speed measurement value d1 of the front virtual object detected by the first camera in the video dark box by the camera image reading module.

[0153] Step S412: determining whether the error between the speed true value C1 and the speed true value D1 is within a first preset speed error threshold range; if not, returning to step S411; if yes, executing step S413.

[0154] Step S413: determining whether the error between the speed measurement value c1 and the speed measurement value d1 is within a second preset speed error threshold range; if not, adjusting the pitch angle of the first camera and returning to step S413; if yes, ending the fine calibration process.

[0155] In the embodiment, the yaw angle, the pitch angle and the roll angle of the camera in the video dark box are fine calibrated according to the preset fine calibration strategy, which effectively improves the accuracy of the relative position of the camera in the dark box, greatly improves the measurement accuracy of the camera in the video dark box scheme and the accuracy of the hardware-in-the-loop function test, and solves the problem that the camera in the video dark box cannot be accurately calibrated after calibration.

[0156] On the basis of the above Figures 1 to 4b The step S105 is a fine calibration verification process, specifically including:

[0157] determining whether the error between the lane line fitting coefficient true value collected by the real vehicle true value collection module and the lane line fitting coefficient true value collected by the ideal sensor module is within a first preset fitting coefficient error threshold; if not, readjusting the position of the virtual ego vehicle in the virtual lane in the virtual scene, and determining whether the error between the lane line fitting coefficient true value collected by the real vehicle true value collection module again and the lane line fitting coefficient true value collected by the ideal sensor module again is within the first preset fitting coefficient error threshold; if yes, then:

[0158] determining whether the error between the lane line fitting coefficient measurement value of the real lane detected by the second camera obtained by the camera image reading module and the lane line fitting coefficient measurement value of the virtual lane detected by the first camera obtained by the camera image reading module is within a second preset fitting coefficient error threshold range; if not, readjusting the jig according to the preset fine calibration strategy; if yes, then:

[0159] determining whether the error between the distance true value of the real ego vehicle from the real object target in front obtained by the real vehicle true value acquisition module and the distance true value of the virtual ego vehicle from the virtual object target in front obtained by the ideal sensor module is within the first preset distance error threshold range; if not, readjusting the positions of the ego vehicle and the preceding vehicle in the virtual scene, and determining whether the error between the distance true value of the real ego vehicle from the real object target in front obtained by the real vehicle true value acquisition module again and the distance true value of the virtual ego vehicle from the virtual object target in front obtained by the ideal sensor module again is within the first preset distance error threshold range; if yes, then:

[0160] determining whether the error between the distance measurement value of the real ego vehicle from the real object target in front detected by the second camera obtained by the camera image reading module and the distance measurement value of the virtual ego vehicle from the virtual object target in front detected by the first camera in the video dark box obtained by the camera image reading module is within the second preset distance error threshold range; if not, readjusting the fixture according to the preset fine mark adjustment strategy; if yes, then determining that the fine mark verification result is passed, and ending the fine mark verification process.

[0161] determining whether the error between the speed true value of the real object target in front obtained by the real vehicle true value acquisition module and the speed true value of the virtual object target in front obtained by the ideal sensor module is within the first preset speed error threshold range; if not, readjusting the speeds of the ego vehicle and the preceding vehicle in the virtual scene, and determining whether the error between the speed true value of the real object target in front obtained by the real vehicle true value acquisition module again and the speed true value of the virtual object target in front obtained by the ideal sensor module again is within the first preset speed error threshold range; if yes, then:

[0162] determining whether the error between the speed measurement value of the real object target in front detected by the second camera obtained by the camera image reading module and the speed measurement value of the virtual object target in front detected by the first camera in the video dark box obtained by the camera image reading module is within the second preset speed error threshold range; if not, readjusting the fixture according to the preset fine mark adjustment strategy; if yes, then determining that the fine mark verification result is passed, and ending the fine mark verification process.

[0163] One specific implementation of the fine mark verification is shown in Figure 4c . Figure 4c An embodiment six of a fine mark verification method based on a video dark box in a driving assistance system provided in the present application is shown in Figure 4c , and the fine mark verification process is executed as follows:

[0164] Step S461: determining whether the error between the lane line fitting coefficient true value collected by the real vehicle true value collection module and the lane line fitting coefficient true value collected by the ideal sensor module is within the first preset fitting coefficient error threshold; if not, readjusting the position of the virtual ego vehicle in the virtual lane in the virtual scene and returning to step S461; if yes, executing step S462.

[0165] Step S462: determining whether the error between the lane line fitting coefficient measured value of the real lane line detected by the second camera detected by the camera image reading module and the lane line fitting coefficient measured value of the virtual lane line detected by the first camera detected by the camera image reading module is within the second preset fitting coefficient error threshold range; if not, returning to the fine label adjustment process and executing step S401; if yes, executing step S463.

[0166] Step S463: determining whether the error between the distance true value A1 and the distance true value B1 is within the first preset distance error threshold range; if not, readjusting the positions of the ego vehicle and the preceding vehicle in the virtual scene and returning to step S463; if yes, executing step S464.

[0167] Step S464: determining whether the error between the distance measured value a1 and the distance measured value b1 is within the second preset distance error threshold range; if not, returning to the fine label adjustment process and executing step S401; if yes, executing step S465.

[0168] Step S465: determining whether the error between the speed true value C1 and the speed true value D1 is within the first preset speed error threshold range; if not, readjusting the speeds of the ego vehicle and the preceding vehicle in the virtual scene and returning to step S465; if yes, executing step S466.

[0169] Step S466: determining whether the error between the speed measured value c1 and the speed measured value d1 is within the second preset speed error threshold range; if not, returning to the fine label adjustment process and executing step S401; if yes, determining that the fine label verification result is passed, and ending the fine label verification process.

[0170] The following is an embodiment of the device of the present application, which can be used to execute the method embodiment of the present application. For details not disclosed in the device embodiment of the present application, please refer to the method embodiment of the present application.

[0171] Figure 5 The following is an embodiment of the device of the present application, which can be used to execute the method embodiment of the present application. For details not disclosed in the device embodiment of the present application, please refer to the method embodiment of the present application. Figure 5As shown, the camera calibration and verification system 50 based on the video dark box includes a camera image reading module 51, an actual vehicle true value collection module 52, an ideal sensor module 53, a camera calibration module 54, an execution module 55, an actual vehicle lane centering and keeping module 56, an ideal lane centering and keeping module 57, a checkerboard target 58, and a bullseye target 59.

[0172] The camera calibration module 54 is configured to store a preset coarse calibration adjustment strategy, perform static calibration and dynamic calibration, and obtain static calibration results and dynamic calibration results, and obtain a yaw angle calibration deviation angle, a pitch angle calibration deviation angle, and a roll angle calibration deviation angle after calibration.

[0173] The execution module 55 is configured to store a pre-adjustment strategy and a fine calibration adjustment strategy, perform corresponding adjustment processing on the first camera in the video dark box according to the preset pre-adjustment strategy, perform coarse calibration adjustment processing on the yaw angle, the pitch angle, and the roll angle of the first camera by adjusting the jig in the video dark box according to the preset coarse calibration adjustment strategy, and perform coarse calibration verification, perform fine calibration adjustment processing on the yaw angle, the pitch angle, and the roll angle of the first camera by adjusting the jig according to the preset fine calibration adjustment strategy after determining that the coarse calibration verification is passed, and perform fine calibration verification, and perform a new round of coarse calibration verification after determining that the fine calibration verification is passed.

[0174] The camera calibration and verification system based on the video dark box provided by the embodiments of the present application can implement the technical solutions shown in the method embodiments, and the implementation principles and beneficial effects are similar, which will not be described here in detail.

[0175] In a possible implementation, the execution module 55 is specifically configured to obtain a preset distance between the first camera and the screen according to the field of view angle and the external parameters of the first camera in the video dark box, and according to the field of view angle and the external parameters of the third camera preset in the scene simulation software, and the information of the screen in the video dark box, and adjust the distance between the first camera and the screen to the preset distance.

[0176] The execution module 55 is specifically configured to determine the position of the lens according to the parameters of the lens in the video dark box and the information of the screen in the video dark box, and coarsely adjust the degrees of freedom of the first camera and the jig in the video dark box in the up-down, front-back, and left-right three directions according to the position of the lens.

[0177] The execution module 55 is specifically configured to finely adjust the degrees of freedom of the first camera in the up-down and left-right two directions according to the camera image reading module, the bullseye target, and the laser pen, so that the center of the adjusted first camera coincides with the bullseye displayed on the screen in the up-down and left-right two directions.

[0178] The bullseye target 59 is used to adjust the horizontal and vertical positions of the first camera and the clamp to be on the same horizontal and vertical lines as the bullseye displayed on the screen.

[0179] The calibration and verification system based on the video dark box in the driving assistance system provided by the embodiments of the present application can execute the technical solutions shown in the method embodiments, and the implementation principles and beneficial effects are similar, which will not be described here in detail.

[0180] In a possible implementation, the execution module 55 is specifically configured to use the level tool to coarsely adjust the roll degree of freedom of the first camera by adjusting the clamp.

[0181] The camera calibration module 54 is specifically configured to obtain the first yaw angle calibration deviation angle, the first pitch angle calibration deviation angle and the first roll angle calibration deviation angle of the first camera.

[0182] If the camera calibration module 54 determines that any of the first yaw angle calibration deviation angle, the first pitch angle calibration deviation angle and the first roll angle calibration deviation angle is not within the corresponding first threshold range, the execution module 55 is specifically configured to adjust the first camera by adjusting the clamp, so that the second yaw angle calibration deviation angle, the second pitch angle calibration deviation angle and the second roll angle calibration deviation angle of the adjusted first camera are all within the corresponding first threshold range.

[0183] The camera calibration module 54 uses the checkerboard target 58 to respectively perform static calibration on the second yaw angle, the second pitch angle and the second roll angle of the first camera; and the execution module 55 is specifically configured to adjust the first camera by adjusting the clamp according to the static calibration results obtained by the camera calibration module, so that the third yaw angle calibration deviation angle, the third pitch angle calibration deviation angle and the third roll angle calibration deviation angle of the first camera reacquired by the camera calibration module after the adjustment are all within the corresponding second threshold range.

[0184] The camera calibration module 54 respectively performs dynamic calibration on the third yaw angle, the third pitch angle and the third roll angle of the first camera; and the execution module 55 is specifically configured to adjust the first camera by adjusting the clamp according to the obtained dynamic calibration results, so that the fourth yaw angle calibration deviation angle, the fourth pitch angle calibration deviation angle and the fourth roll angle calibration deviation angle of the adjusted first camera are all within the corresponding third threshold range.

[0185] The checkerboard target 58 is used for static calibration in coarse calibration and static calibration in coarse verification.

[0186] The camera calibration and verification system based on the video dark box provided by the embodiments of the present application can execute the technical solutions shown in the method embodiments, and the implementation principles and beneficial effects are similar, which will not be repeated here.

[0187] In a possible implementation, the execution module 55 is specifically configured to, when the camera calibration module 54 determines that the calibration deviation angle of the roll angle of the first camera after the rough calibration adjustment is not within the preset roll angle deviation threshold range, finely adjust the roll angle through the level, so that the calibration deviation angle of the roll angle obtained by the camera calibration module after the adjustment is within the first preset roll angle deviation threshold range, and the deviation value of the roll angle measurement value from the horizontal reference value is within the second preset roll angle deviation threshold range.

[0188] The real vehicle true value collection module 52 is configured to, when the real vehicle carrying the second camera drives on the center of the real lane, obtain the lane line fitting coefficient true value on the real lane.

[0189] The ideal sensor module 53 is configured to, when the virtual vehicle carrying the third camera drives on the center of the virtual lane, obtain the lane line fitting coefficient true value on the virtual lane detected by the third camera.

[0190] The camera image reading module 51 is configured to, when the real vehicle carrying the second camera drives on the center of the real lane, obtain the lane line fitting coefficient measurement value on the real lane detected by the second camera.

[0191] The camera image reading module 51 is further configured to, when the virtual vehicle carrying the third camera drives on the center of the virtual lane, obtain the virtual lane line fitting coefficient measurement value on the dark box screen detected by the first camera in the video dark box.

[0192] The real vehicle lane center keeping module 56 is configured to keep the real vehicle carrying the second camera always driving on the center of the real lane.

[0193] The ideal lane center keeping module 57 is configured to keep the virtual vehicle carrying the third camera always driving on the center of the virtual lane.

[0194] The execution module 55 is specifically further configured to adjust the yaw angle of the first camera when the error between the lane line fitting coefficient true value collected by the real vehicle true value collection module 52 and the lane line fitting coefficient true value collected by the ideal sensor module 53 is within the first preset fitting coefficient error threshold, and the error between the lane line fitting coefficient measurement value of the real lane line detected by the second camera of the camera image reading module 51 and the lane line fitting coefficient measurement value of the virtual lane line in the video dark box screen detected by the first camera of the camera image reading module 51 exceeds the second preset fitting coefficient error threshold range, so that the error between the two updated measurement values is within the second preset fitting coefficient error threshold range.

[0195] The real vehicle true value collection module 52 is further configured to acquire the distance true value of the real vehicle from the front real object when the real vehicle carrying the second camera is stationary in the real lane center, and the front real object is also stationary in the real lane center.

[0196] The ideal sensor module 53 is further configured to acquire the distance true value of the virtual vehicle from the front virtual object detected by the third camera when the virtual vehicle carrying the third camera is stationary in the virtual lane center displayed on the video dark box screen, and the front virtual object is also stationary in the virtual lane center.

[0197] The camera image reading module 51 is configured to acquire the distance measurement value of the real vehicle from the front real object detected by the second camera when the real vehicle carrying the second camera is stationary in the real lane center, and the front real object is also stationary in the real lane center.

[0198] The camera image reading module 51 is further configured to acquire the distance measurement value of the virtual vehicle from the front virtual object on the video dark box screen detected by the first camera when the virtual vehicle carrying the third camera is stationary in the virtual lane center displayed on the video dark box screen, and the front virtual object is also stationary in the virtual lane center.

[0199] The execution module 55 is specifically configured to adjust the pitch angle of the first camera when the error between the distance true value of the real object target in front collected by the real car true value collection module 52 and the distance true value of the virtual object target in front collected by the ideal sensor module 53 is within the first preset distance error threshold, and the error between the distance measurement value of the real car in front detected by the second camera of the camera reading image module 51 and the distance measurement value of the virtual car in front detected by the first camera of the camera reading image module 51 is greater than the second preset distance error threshold range, so that the error between the updated distance measurement value of the real car in front and the distance measurement value of the virtual car in front in the adjusted first camera is within the second preset distance error threshold range.

[0200] The real car true value collection module 52 is further configured to acquire the speed true value of the front real object target on the real road detected by the second camera when the real car carrying the second camera is uniformly driven at a set speed on the real lane center, and the front real object target is also uniformly driven at a specific speed on the real lane center.

[0201] The ideal sensor module 53 is further configured to acquire the speed true value of the front virtual object target detected by the third camera when the virtual car carrying the third camera is uniformly driven at a set speed on the virtual lane center displayed on the video dark box screen, and the front virtual object target is also uniformly driven at a specific speed on the virtual lane center.

[0202] The camera reading image module 51 is configured to acquire the speed measurement value of the front real object target on the real lane detected by the second camera when the real car carrying the second camera is uniformly driven at a set speed on the real lane center, and the front real object target is also uniformly driven at a specific speed on the real lane center.

[0203] The camera reading image module 51 is further configured to acquire the speed measurement value of the front virtual object target on the dark box screen detected by the first camera when the virtual car carrying the third camera is uniformly driven at a set speed on the virtual lane center displayed on the video dark box screen, and the front virtual object target is also uniformly driven at a specific speed on the virtual lane center.

[0204] The execution module 55 is specifically configured to adjust the pitch angle of the first camera when the error between the speed true value of the front real object collected by the real vehicle true value collection module 52 and the speed true value of the front virtual object collected by the ideal sensor module 53 is within the first preset speed error threshold, and when the error between the speed measurement value of the front real object detected by the second camera obtained by the camera image reading module 51 and the speed measurement value of the front virtual object in the video dark box screen detected by the first camera obtained by the camera image reading module 51 exceeds the second preset speed error threshold range, so that the error between the updated speed measurement value of the front real object and the speed measurement value of the front virtual object is within the second preset speed error threshold range.

[0205] In the coarse calibration verification process:

[0206] The camera calibration module and the checkerboard target are used to respectively re-carry out static calibration on the third yaw angle, the third pitch angle and the third roll angle of the first camera, and the camera calibration module is used to obtain the fifth yaw angle calibration deviation angle, the fifth pitch angle calibration deviation angle and the fifth roll angle calibration deviation angle after static calibration.

[0207] The camera calibration module and the dynamic traffic scene played on the screen in the video dark box are used to respectively carry out dynamic calibration on the third yaw angle, the third pitch angle and the third roll angle of the first camera, and the camera calibration module is used to obtain the sixth yaw angle calibration deviation angle, the sixth pitch angle calibration deviation angle and the sixth roll angle calibration deviation angle after dynamic calibration.

[0208] In the fine calibration verification process:

[0209] The camera image reading module is used to obtain the lane line fitting coefficient measurement value of the real lane detected by the second camera, and obtain the lane line fitting coefficient measurement value of the real lane detected by the second camera. According to the preset fine calibration algorithm, the execution module is used to judge the error between the two and the second preset fitting coefficient error threshold range.

[0210] The camera image reading module is also used to obtain the distance measurement value of the real vehicle from the front real object detected by the second camera, and obtain the distance measurement value of the virtual vehicle from the front virtual object detected by the first camera in the video dark box. And according to the preset fine calibration algorithm, the execution module is used to judge the error between the two and the second preset distance error threshold range.

[0211] The camera reading module can also obtain a speed measurement of the real object in front detected by the second camera, and obtain a speed measurement of the virtual object in front detected by the first camera in the video dark box. According to a preset precision calibration algorithm, the execution module is used to judge the error between the two and the second preset speed error threshold range.

[0212] The camera calibration verification system based on the video dark box provided by the embodiments of the present application can execute the technical solutions shown in the method embodiments, and the implementation principles and beneficial effects are similar, which will not be described here.

[0213] Those skilled in the art can understand that all or part of the steps of the above-mentioned method embodiments can be completed by program instruction related hardware. The foregoing program can be stored in a computer readable storage medium. When the program is executed, the steps of the above-mentioned method embodiments are executed; and the foregoing storage medium includes ROM, RAM, magnetic disc or optical disc and various storage medium capable of storing program codes.

[0214] Finally, it should be pointed out that: the above embodiments are only used to illustrate the technical solutions of the present application, and not to limit them; although the present application has been described in detail with reference to the foregoing embodiments, those skilled in the art should understand that: it can still modify the technical solutions recorded in the foregoing embodiments, or make equivalent replacement for part or all of the technical features; and these modifications or replacements do not make the essence of the corresponding technical solutions deviate from the scope of the technical solutions of the embodiments of the present application.

Claims

1. A camera calibration verification method based on video dark box, characterized in that, The method comprises the following steps: According to the preset front adjustment strategy, the first camera in the video dark box is adjusted correspondingly; According to the preset rough adjustment strategy, the yaw angle, pitch angle and roll angle of the first camera are adjusted and checked respectively by adjusting the clamp in the video dark box; After determining that the rough check is passed, according to the preset fine adjustment strategy, the yaw angle, pitch angle and roll angle of the first camera are adjusted and checked respectively by adjusting the clamp; After determining that the fine check is passed, a new round of rough check is performed again; The fine adjustment comprises: When the roll angle deviation angle obtained by the camera calibration module after adjustment is within the first preset roll angle deviation threshold range, and the deviation value of the roll angle measurement value and the horizontal reference value is within the second preset roll angle deviation threshold range, the roll degree of freedom is fixed; The real vehicle equipped with the second camera is set to drive at a preset speed on the real lane, and the real vehicle lane centering module is started to ensure that the real vehicle always drives in the lane center; The lane line fitting coefficient true value collected on the real lane is obtained by using the real vehicle true value acquisition module, and the lane line fitting coefficient measurement value detected by the second camera on the real lane is obtained by using the camera image reading module; The virtual vehicle equipped with the third camera is set to drive at the preset speed on the virtual lane displayed on the screen in the video dark box, and the ideal lane centering module is started to ensure that the virtual vehicle always drives in the lane center; The lane line fitting coefficient true value collected on the virtual lane displayed on the screen in the video dark box is obtained by using the ideal sensor module, and the lane line fitting coefficient measurement value detected by the first camera in the video dark box is obtained by using the camera image reading module; When the error between the lane line fitting coefficient true value collected by the real vehicle true value acquisition module and the lane line fitting coefficient true value collected by the ideal sensor module is within the first preset fitting coefficient error threshold, and the error between the lane line fitting coefficient measurement value of the real lane detected by the second camera obtained by the camera image reading module and the lane line fitting coefficient measurement value of the virtual lane in the screen of the video dark box detected by the first camera obtained by the camera image reading module exceeds the second preset fitting coefficient error threshold range, the yaw angle of the first camera is adjusted, so that the error between the two updated measurement values is within the second preset fitting coefficient error threshold range; The real vehicle equipped with the second camera is set to be static in the real lane center, and the front real object is also static in the real lane center, the distance true value of the real vehicle from the front real object is obtained by using the real vehicle true value acquisition module, and the distance measurement value of the real vehicle from the front real object detected by the second camera is obtained by using the camera image reading module; The third camera is arranged on the virtual vehicle, and the virtual vehicle is arranged on the screen display of the virtual lane in the video dark box, and the front virtual target is also arranged on the screen display of the virtual lane in the video dark box; the ideal sensor module is used to obtain the distance true value of the virtual vehicle from the front virtual target collected on the screen display of the dark box; and the camera image reading module is used to obtain the distance measurement value of the virtual vehicle from the front virtual target detected by the first camera in the video dark box. When the distance error between the distance true value of the virtual vehicle from the front virtual target obtained by the ideal sensor module and the distance true value of the real vehicle from the front real target obtained by the real vehicle true value acquisition module is within a first preset distance error threshold range, and the error between the distance measurement value of the real vehicle from the front real target detected by the second camera and the distance measurement value of the virtual vehicle from the front virtual target detected by the first camera is greater than a second preset distance error threshold range, the pitch angle of the first camera is adjusted so that the error between the updated distance measurement value of the real vehicle from the front real target and the distance measurement value of the virtual vehicle from the front virtual target is within the second preset distance error threshold range. The second camera is arranged on the real vehicle, and the real vehicle is arranged on the real lane at a constant speed, and the front real target is also arranged on the real lane at a constant speed; the real vehicle true value acquisition module is used to obtain the speed true value of the front real target; and the camera image reading module is used to obtain the speed measurement value of the front real target detected by the second camera. The third camera is arranged on the virtual vehicle, and the virtual vehicle is arranged on the screen display of the virtual lane in the video dark box, and the front virtual target is also arranged on the screen display of the virtual lane in the video dark box; the ideal sensor module is used to obtain the speed true value of the front virtual target; and the camera image reading module is used to obtain the speed measurement value of the front virtual target detected by the first camera in the video dark box. When the speed of the virtual vehicle and the speed of the front virtual target are adjusted to be consistent with the working condition of the real vehicle in the scene simulation software, and the error between the speed true value of the front virtual target obtained by the ideal sensor module and the speed true value of the front real target obtained by the real vehicle true value acquisition module is within a first preset speed error threshold range, and the error between the speed measurement value of the front real target detected by the second camera and the speed measurement value of the front virtual target detected by the first camera is greater than a second preset speed error threshold range, the pitch angle of the first camera is adjusted so that the error between the updated speed measurement value of the front real target and the speed measurement value of the front virtual target is within the second preset speed error threshold range.

2. The method of claim 1, wherein, The first camera in the video dark box is adjusted according to a preset front adjustment strategy, including: A screen position is preset, and a field of view angle and an external parameter of a third camera in a scene simulation software are preset; A preset distance between the first camera and the screen in the video dark box is obtained according to the field of view angle and the external parameter of the first camera, according to the preset field of view angle and the external parameter of the third camera in the scene simulation software, and according to information of the screen in the video dark box, and the distance between the first camera and the screen is adjusted to the preset distance; The position of the lens in the video dark box is determined according to parameters of the lens and information of the screen in the video dark box, and the degrees of freedom of the first camera and a clamp in the video dark box in up-down, front-back and left-right directions are coarsely adjusted according to the position of the lens. The horizontal and vertical degrees of freedom of the first camera are finely adjusted by a camera image reading module, a target center target and a laser pen, so that the center of the adjusted first camera coincides with the target center displayed on the screen in the horizontal and vertical directions.

3. The method of claim 1, wherein, The yaw angle, pitch angle and roll angle of the first camera are coarsely adjusted by adjusting the clamp in the video dark box according to a preset coarse adjustment strategy, including: The roll degree of freedom of the first camera is coarsely adjusted by adjusting the clamp by using a level tool; The first yaw angle calibration deviation angle, the first pitch angle calibration deviation angle and the first roll angle calibration deviation angle of the first camera are obtained by using a camera calibration module; If any of the first yaw angle calibration deviation angle, the first pitch angle calibration deviation angle and the first roll angle calibration deviation angle is not within the corresponding first threshold range, the first camera is adjusted by adjusting the clamp, so that the second yaw angle calibration deviation angle, the second pitch angle calibration deviation angle and the second roll angle calibration deviation angle of the adjusted first camera are all within the corresponding first threshold range; The second yaw angle, the second pitch angle and the second roll angle of the first camera are statically calibrated by using a camera calibration module and a checkerboard target, and the first camera is adjusted by adjusting the clamp according to the static calibration result obtained by the camera calibration module, so that the third yaw angle calibration deviation angle, the third pitch angle calibration deviation angle and the third roll angle calibration deviation angle of the first camera reacquired by the camera calibration module after adjustment are all within the corresponding second threshold range; The third roll angle, the third pitch angle and the third roll angle of the first camera are dynamically calibrated respectively by using the camera calibration module and the dynamic traffic scene played on the screen in the video dark box, and the first camera is adjusted by adjusting the jig according to the dynamic calibration result obtained by the camera calibration module, so that the fourth roll angle, the fourth pitch angle and the fourth roll angle of the first camera reacquired by the camera calibration module after adjustment are all within the corresponding third threshold range; The coarse calibration verification includes: The third roll angle, the third pitch angle and the third roll angle of the first camera are dynamically calibrated respectively by using the camera calibration module and the dynamic traffic scene played on the screen in the video dark box, and the first camera is adjusted by adjusting the jig according to the dynamic calibration result obtained by the camera calibration module, so that the fourth roll angle, the fourth pitch angle and the fourth roll angle of the first camera reacquired by the camera calibration module after adjustment are all within the corresponding third threshold range; The fifth roll angle, the fifth pitch angle and the fifth roll angle of the first camera are dynamically calibrated respectively by using the camera calibration module and the dynamic traffic scene played on the screen in the video dark box, and the dynamic calibration result is obtained by the camera calibration module; The sixth roll angle, the sixth pitch angle and the sixth roll angle of the first camera are dynamically calibrated respectively by using the camera calibration module and the dynamic traffic scene played on the screen in the video dark box, and the dynamic calibration result is obtained by the camera calibration module; The sixth roll angle, the sixth pitch angle and the sixth roll angle of the first camera are dynamically calibrated respectively by using the camera calibration module and the dynamic traffic scene played on the screen in the video dark box, and the dynamic calibration result is obtained by the camera calibration module; 4. The method according to any one of claims 1 to 3, characterized in that, The fine calibration verification includes: The error between the lane line fitting coefficient true value collected by the real vehicle true value collection module and the lane line fitting coefficient true value collected by the ideal sensor module is judged whether it is within the first preset fitting coefficient error threshold; if not, the position of the virtual ego vehicle in the virtual lane in the virtual scene is re-adjusted, and the error between the lane line fitting coefficient true value collected by the real vehicle true value collection module and the lane line fitting coefficient true value collected by the ideal sensor module is judged whether it is within the first preset fitting coefficient error threshold; if yes, then: The error between the lane line fitting coefficient measured value of the real lane detected by the second camera and the lane line fitting coefficient measured value of the virtual lane detected by the first camera is judged whether it is within the second preset fitting coefficient error threshold range; if not, the jig is re-adjusted according to the preset fine calibration adjustment strategy; if yes, then: determining whether the error between the distance true value of the real car from the real target in front obtained by the real car true value acquisition module and the distance true value of the virtual car from the virtual target in front obtained by the ideal sensor module is within the first preset distance error threshold range; if not, readjusting the positions of the car and the front car in the virtual scene, and determining whether the error between the distance true value of the real car from the real target in front obtained by the real car true value acquisition module again and the distance true value of the virtual car from the virtual target in front obtained by the ideal sensor module again is within the first preset distance error threshold range; if yes, then: determining whether the error between the distance measurement value of the real car from the real target in front detected by the second camera obtained by the camera image reading module and the distance measurement value of the virtual car from the virtual target in front detected by the first camera in the video dark box obtained by the camera image reading module is within the second preset distance error threshold range; if not, then readjusting the jig according to the preset fine mark adjustment strategy; if yes, then determining that the fine mark verification result passes, and ending the fine mark verification process. determining whether the error between the speed true value of the real target in front obtained by the real car true value acquisition module and the speed true value of the virtual target in front obtained by the ideal sensor module is within the first preset speed error threshold range; if not, readjusting the speeds of the car and the front car in the virtual scene, and determining whether the error between the speed true value of the real target in front obtained by the real car true value acquisition module again and the speed true value of the virtual target in front obtained by the ideal sensor module again is within the first preset speed error threshold range; if yes, then: determining whether the error between the speed measurement value of the real target in front detected by the second camera obtained by the camera image reading module and the speed measurement value of the virtual target in front detected by the first camera in the video dark box obtained by the camera image reading module is within the second preset speed error threshold range; if not, then readjusting the jig according to the preset fine mark adjustment strategy; if yes, then determining that the fine mark verification result passes, and ending the fine mark verification process.

5. A video- darkroom-based camera calibration verification system, characterized by, The camera calibration verification system based on the video dark box is used to implement the camera calibration verification method based on the video dark box in any one of claims 1-4, and comprises a camera image reading module, a real car true value acquisition module, an ideal sensor module, a camera calibration module, an execution module, a real car lane centering and keeping module, an ideal lane centering and keeping module, a checkerboard target and a target center target. The camera calibration module is used to store a preset coarse mark adjustment strategy, perform static calibration and dynamic calibration, obtain static calibration results and dynamic calibration results, and obtain calibrated yaw angle calibration deviation angle, pitch angle calibration deviation angle and roll angle calibration deviation angle. The execution module is used to: store a pre-adjustment strategy and a fine mark adjustment strategy; perform corresponding adjustment processing on the first camera in the video dark box according to the preset pre-adjustment strategy; perform coarse mark adjustment processing on the yaw angle, the pitch angle and the roll angle of the first camera respectively by adjusting the jig in the video dark box according to the preset coarse mark adjustment strategy, and perform coarse mark verification. After determining that the coarse calibration passes the verification, according to a preset fine calibration adjustment strategy, the yaw angle, the pitch angle and the roll angle of the first camera are adjusted and fine calibration verification is performed by adjusting the clamp. After determining that the fine calibration passes the verification, a new round of coarse calibration verification is performed.

6. The system of claim 5, wherein, The execution module is specifically configured to: According to the field of view angle and the external parameter of the first camera in the video dark box, and according to the field of view angle and the external parameter of the third camera preset in the scene simulation software and the information of the screen in the video dark box, a preset distance between the first camera and the screen is obtained, and the distance between the first camera and the screen is adjusted to the preset distance. According to the parameters of the lens in the video dark box and the information of the screen in the video dark box, the position of the lens is determined, and the degrees of freedom of the first camera and the clamp in the video dark box in the up-down, front-back and left-right directions are coarsely adjusted according to the position of the lens. According to the camera reading module, the target center target and the laser pen, the degrees of freedom of the first camera in the up-down and left-right directions are finely adjusted, so that the center of the adjusted first camera coincides with the target center displayed on the screen in the up-down and left-right directions. The target center target is used to adjust the up-down and left-right positions of the first camera and the clamp to be on the same horizontal line and the same vertical line as the target center displayed on the screen.

7. The system of claim 5, wherein, The execution module is specifically configured to: The degrees of freedom of the roll of the first camera are coarsely adjusted by adjusting the clamp by using the level tool. The camera calibration module is specifically configured to obtain the first yaw angle calibration deviation angle, the first pitch angle calibration deviation angle and the first roll angle calibration deviation angle of the first camera. If any of the first yaw angle calibration deviation angle, the first pitch angle calibration deviation angle and the first roll angle calibration deviation angle determined by the camera calibration module is not within the corresponding first threshold range, the execution module is specifically configured to adjust the first camera by adjusting the clamp, so that the second yaw angle calibration deviation angle, the second pitch angle calibration deviation angle and the second roll angle calibration deviation angle of the adjusted first camera are all within the corresponding first threshold range. The camera calibration module uses the checkerboard target to respectively calibrate the second yaw angle, the second pitch angle and the second roll angle of the first camera, and the execution module is specifically configured to adjust the first camera by adjusting the clamp according to the static calibration result obtained by the camera calibration module, so that the third yaw angle calibration deviation angle, the third pitch angle calibration deviation angle and the third roll angle calibration deviation angle of the first camera reacquired by the camera calibration module after the adjustment are all within the corresponding second threshold range. The camera calibration module respectively dynamically calibrates a third yaw angle, a third pitch angle and a third roll angle of the first camera; and the execution module is specifically configured to: according to the obtained dynamic calibration result, adjusting the first camera by adjusting the clamp, so that a fourth yaw angle calibration deviation angle, a fourth pitch angle calibration deviation angle and a fourth roll angle calibration deviation angle of the adjusted first camera are all within the corresponding third threshold range. The checkerboard target is used for static calibration in coarse adjustment and static calibration in coarse calibration.

8. The system of any one of claims 5 to 7, wherein, The execution module is specifically configured to: when the camera calibration module determines that a calibration deviation angle of a roll angle of the first camera after coarse adjustment is not within a preset roll angle deviation threshold range, adjusting the roll angle by using a level, so that a calibration deviation angle of the roll angle obtained by using the camera calibration module after adjustment is within a first preset roll angle deviation threshold range, and a deviation value of the roll angle measurement value from the horizontal reference value is within a second preset roll angle deviation threshold range; The real vehicle true value acquisition module is configured to acquire a lane line fitting coefficient true value on a real lane when a real ego vehicle carrying the second camera drives on a real lane center; The ideal sensor module is configured to acquire a lane line fitting coefficient true value on a virtual lane detected by the third camera when a virtual ego vehicle carrying the third camera drives on a virtual lane center; The camera image reading module is configured to acquire a lane line fitting coefficient measurement value on a real lane detected by the second camera when the real ego vehicle carrying the second camera drives on the real lane center; The camera image reading module is further configured to acquire a virtual lane line fitting coefficient measurement value on a screen in the video dark box detected by the first camera when the virtual ego vehicle carrying the third camera drives on the virtual lane center; The real vehicle lane center keeping module is configured to keep the real ego vehicle carrying the second camera always driving on the real lane center; The ideal lane center keeping module is configured to keep the virtual ego vehicle carrying the third camera always driving on the virtual lane center; The execution module is specifically further configured to: when an error between the lane line fitting coefficient true value acquired by using the real vehicle true value acquisition module and the lane line fitting coefficient true value acquired by using the ideal sensor module is within a first preset fitting coefficient error threshold, and an error between a lane line fitting coefficient measurement value on a real lane detected by the second camera and a lane line fitting coefficient measurement value on a virtual lane in the video dark box detected by the first camera exceeds a second preset fitting coefficient error threshold range, adjusting a yaw angle of the first camera, so that the error between the two updated measurement values is within the second preset fitting coefficient error threshold range. The real vehicle true value collection module is further configured to acquire a distance true value of the real vehicle from the front real object when the real vehicle carrying the second camera is static at a real lane center and the front real object is also static at the real lane center; The ideal sensor module is further configured to acquire a distance true value of the virtual vehicle from the front virtual object detected by the third camera when the virtual vehicle carrying the third camera is static at a virtual lane center displayed on the video dark box screen and the front virtual object is also static at the virtual lane center; The camera image reading module is configured to acquire a distance measurement value of the real vehicle from the front real object detected by the second camera when the real vehicle carrying the second camera is static at a real lane center and the front real object is also static at the real lane center; The camera image reading module is further configured to acquire a distance measurement value of the virtual vehicle from the front virtual object detected by the first camera on the video dark box screen when the virtual vehicle carrying the third camera is static at a virtual lane center displayed on the video dark box screen and the front virtual object is also static at the virtual lane center; The execution module is specifically configured to adjust the pitch angle of the first camera when an error between the distance true value of the real vehicle from the front real object collected by the real vehicle true value collection module and the distance true value of the virtual vehicle from the front virtual object collected by the ideal sensor module is within a first preset distance error threshold, and when an error between the distance measurement value of the real vehicle from the front real object detected by the second camera acquired by the camera image reading module and the distance measurement value of the virtual vehicle from the front virtual object detected by the first camera acquired by the camera image reading module exceeds a second preset distance error threshold range, so that an error between an updated distance measurement value of the real vehicle from the front real object in the adjusted first camera and the distance measurement value of the virtual vehicle from the front virtual object is within the second preset distance error threshold range; The real vehicle true value collection module is further configured to acquire a speed true value of the front real object on the real road detected by the second camera when the real vehicle carrying the second camera is uniformly driven at a set speed on the real lane center and the front real object is also uniformly driven at a specific speed on the real lane center; The ideal sensor module is further configured to acquire a speed true value of the front virtual object detected by the third camera when the virtual vehicle carrying the third camera is uniformly driven at a set speed on the virtual lane center displayed on the video dark box screen and the front virtual object is also uniformly driven at a specific speed on the virtual lane center; The camera image reading module is configured to acquire a speed measurement value of the front real object on the real lane detected by the second camera when the real vehicle carrying the second camera is uniformly driven at a set speed on the real lane center and the front real object is also uniformly driven at a specific speed on the real lane center; The camera image reading module is further configured to acquire a speed measurement value of the front virtual target on the dark box screen detected by the first camera when the virtual car carrying the third camera is uniformly driven at a set speed on the virtual lane center displayed on the video dark box screen, and the front virtual target is also uniformly driven at a specific speed on the virtual lane center. The execution module is specifically configured to adjust the pitch angle of the first camera when an error between a speed true value of the front real target collected by the real car true value collection module and a speed true value of the front virtual target collected by the ideal sensor module is within a first preset speed error threshold, and an error between a speed measurement value of the front real target detected by the second camera acquired by the camera image reading module and a speed measurement value of the front virtual target in the video dark box screen detected by the first camera acquired by the camera image reading module exceeds a second preset speed error threshold range, so that an error between the updated speed measurement value of the front real target and the speed measurement value of the front virtual target is within the second preset speed error threshold range.

Images