A multi-angle detection method, system and storage medium for a steering knuckle

Abstract

The present invention relates to a multi-angle detection method, system and storage medium for a steering knuckle. The method includes: U1. On the production line of the steering knuckle, based on a surface array industrial camera, multi-angle image data information of the steering knuckle is obtained in real time, and based on a laser scanner, multi-angle point cloud data information of the steering knuckle is obtained in real time, and preprocessing is performed to obtain the preprocessed multi-angle image data information and point cloud data information of the steering knuckle; U2. Based on the preprocessed multi-angle image data information and point cloud data information of the steering knuckle, a time synchronization optimization algorithm is used to optimize the image and point cloud data information of the steering knuckle, and the optimized multi-angle image data information and the optimized multi-angle point cloud data information of the steering knuckle are obtained. The present invention not only solves the problems of unstable and deviated detection of the steering knuckle, but also the entire detection process is convenient, fast and highly accurate, and can further improve the production efficiency of the steering knuckle.

Description

Technical Field

[0001] The present invention relates to the technical field of automotive parts inspection, and particularly to a multi-angle inspection method, system and storage medium for a steering knuckle. Background Art

[0002] With the rapid development of the automotive industry, automotive parts manufacturers are constantly improving new technologies. The introduction of new technologies has improved production efficiency and product quality. With the emphasis on quality first and high-precision requirements, enterprises can remain invincible in the increasingly fierce competition. Therefore, product quality has attracted the key attention of all enterprises in the increasingly fierce competition.

[0003] The steering knuckle is an important safety part of the vehicle. Industry professionals all over the world attach great importance to its safety characteristics. Fully understanding its safety characteristics plays a particularly important role in the inspection of the steering knuckle and should be integrated into the control of the entire manufacturing process. The steering knuckle is a mechanical part that realizes variable-angle power transmission and is used in positions where the direction of the transmission axis needs to be changed. It is the "joint" part of the universal transmission device of the vehicle drive system. If there are any dimensional accuracy and technical problems with the steering knuckle, it will have a great impact on vehicle safety.

[0004] Therefore, in the production process of the steering knuckle, the inspection of the accuracy of the steering knuckle is very important. In the prior art, a sampling manual inspection method is adopted, which cannot comprehensively inspect the steering knuckle, increases the defective rate of the steering knuckle, and at the same time, consumes a long cycle and reduces the production efficiency of the steering knuckle. Summary of the Invention

[0005] In view of the above problems, the present invention provides a multi-angle inspection method, system and storage medium for a steering knuckle, which not only solves the problem of low inspection accuracy of the steering knuckle, but also makes the entire inspection process convenient and fast, and can further improve the production efficiency of the steering knuckle.

[0006] In order to achieve the above object and other related objects, the technical solution provided by the present invention is as follows:

[0007] A multi-angle inspection method for a steering knuckle, the method comprising:

[0008] U1. On the production line of the steering knuckle, based on a surface array industrial camera, multi-angle image data information of the steering knuckle is obtained in real time, and based on a laser scanner, multi-angle point cloud data information of the steering knuckle is obtained in real time, and preprocessing is performed to obtain preprocessed multi-angle image data information and point cloud data information of the steering knuckle;

[0009] U2. Based on the multi - angle image data information and point cloud data information of the pre - processed steering knuckle, use the time - synchronization optimization algorithm to optimize the image and point cloud data information of the steering knuckle, and obtain the multi - angle image data information of the optimized steering knuckle and the multi - angle point cloud data information of the optimized steering knuckle;

[0010] U3. Based on the multi - angle image data information of the optimized steering knuckle and the multi - angle point cloud data information of the optimized steering knuckle, use the multi - angle feature fusion algorithm to fuse the image and point cloud data information of the steering knuckle, and output the multi - angle data information of the fused steering knuckle;

[0011] U4. Based on the multi - angle data information of the fused steering knuckle, use the multi - angle matching algorithm of the steering knuckle to match with the preset standard model of the steering knuckle, and obtain the multi - angle matching degree data information of the steering knuckle;

[0012] U5. Based on the multi - angle matching degree data information of the steering knuckle, construct the evaluation function M of the steering knuckle to evaluate the matching degree of the steering knuckle, and obtain the multi - angle detection data information of the steering knuckle.

[0013] Further, in step U1, the pre - processing includes:

[0014] U11. Denoise, enhance the image, and perform image registration on the multi - angle image data information of the steering knuckle to obtain the multi - angle image data information of the pre - processed steering knuckle;

[0015] U12. Remove outliers, filter, and voxelize the multi - angle point cloud data information of the steering knuckle to obtain the multi - angle point cloud data information of the pre - processed steering knuckle.

[0016] Further, in step U2, the use of the time - synchronization optimization algorithm to optimize the image and point cloud data information of the steering knuckle includes:

[0017] U21. Based on the multi - angle image data information and point cloud data information of the pre - processed steering knuckle, perform timestamp calibration to obtain the multi - angle image data information of the calibrated steering knuckle and the multi - angle point cloud data information of the calibrated steering knuckle;

[0018] U22. Based on the multi - angle image data information of the calibrated steering knuckle and the multi - angle point cloud data information of the calibrated steering knuckle, for each angle of the steering knuckle, establish the image time - synchronization optimization function G of the steering knuckle and the point cloud time - synchronization optimization function H of the steering knuckle,

[0019] ,

[0020] ,

[0021] where T is the sampling period, t is different moments, ζ is a constant parameter, X t is the multi-angle image data information of the calibrated knuckle at time t, Y t is the multi-angle point cloud data information of the calibrated knuckle at time t, X t+ζ is the multi-angle image data information of the calibrated knuckle at time t + ζ, Y t+ζ is the multi-angle point cloud data information of the calibrated knuckle at time t + ζ. Process the multi-angle image data information and point cloud data information of the calibrated knuckle to obtain the processed multi-angle image data information and processed multi-angle point cloud data information of the knuckle;

[0022] U23. Based on the processed multi-angle image data information and the processed multi-angle point cloud data information of the knuckle, synchronize them according to the time series to obtain the optimized multi-angle image data information and optimized multi-angle point cloud data information of the knuckle.

[0023] Further, the synchronization according to the time series is to make the processed multi-angle image data information of the knuckle correspond one by one with the processed multi-angle point cloud data information of the knuckle according to the time points, and eliminate the multi-angle image data information or multi-angle point cloud data information of the knuckle that cannot be corresponded.

[0024] Further, in step U3, the fusion of the image and point cloud data information of the knuckle using the multi-angle feature fusion algorithm includes:

[0025] U31. Input the optimized multi-angle image data information of the knuckle into the trained convolutional neural network model for feature point extraction to obtain the multi-angle image feature point data information of the knuckle;

[0026] U32. Based on the multi-angle image feature point data information of the knuckle and the optimized multi-angle point cloud data information of the knuckle, establish a feature fusion function M,

[0027] ,

[0028] where A is the multi-angle image feature point data information of the knuckle, B is the optimized multi-angle point cloud data information of the knuckle, and α and β are feature fusion parameter matrices;

[0029] U33. Based on the feature fusion function M, obtain the fused multi-angle data information of the knuckle.

[0030] Further, in step U4, the matching of the multi-angle matching algorithm of the knuckle with the preset standard model of the knuckle includes:

[0031] U41. Based on the multi-angle data information of the fused knuckle and the preset standard model of the knuckle, construct a matching matrix of the knuckle to obtain the multi-angle matching matrix data information of the fused knuckle and the matching matrix data information of the standard model;

[0032] U42. Based on the multi-angle matching matrix data information of the fused knuckle and the matching matrix data information of the standard model, establish a multi-angle matching function N of the knuckle,

[0033] ,

[0034] where C is the multi-angle matching matrix data information of the fused knuckle, D is the matching matrix data information of the standard model, γ and δ are transformation matrices, and µ is a matching constant parameter;

[0035] U43. Based on the multi-angle matching function N of the knuckle, obtain the multi-angle matching degree data information of the knuckle.

[0036] Further, the constraint condition of the matching constant parameter µ is:

[0037] .

[0038] Further, in step U5, construct an evaluation function M of the knuckle,

[0039] ,

[0040] where n is the sample size, θ ij is the j-th matching degree data information of the knuckle at the i-th angle, and σ ij is the weight coefficient.

[0041] To achieve the above object and other related objects, the present invention also provides a multi-angle detection system applied to a knuckle, including a computer device, which is programmed or configured to execute the steps of any one of the multi-angle detection methods applied to a knuckle.

[0042] To achieve the above object and other related objects, the present invention also provides a computer-readable storage medium, on which a computer program is stored that is programmed or configured to execute the multi-angle detection method applied to a knuckle according to any one of the above.

[0043] The present invention has the following positive effects:

[0044] 1. The present invention optimizes the image and point cloud data information of the steering knuckle by adopting a time synchronization optimization algorithm, and combines a multi-angle feature fusion algorithm to fuse the image and point cloud data information of the steering knuckle, which can not only accurately master the data information of each angle of the steering knuckle, but also improve the robustness and stability of the algorithm, providing accurate data support for the subsequent detection of the steering knuckle.

[0045] 2. The present invention matches the multi-angle matching algorithm of the steering knuckle with a preset standard model of the steering knuckle, and combines the evaluation function of the steering knuckle to comprehensively evaluate the detection of the steering knuckle, which not only improves the detection accuracy of the steering knuckle, but also does not require manual participation in the whole detection process, reducing labor costs and improving production efficiency.

[0046] 3. The whole process of the steering knuckle detection of the present invention is convenient and fast, which can further improve the production efficiency of the steering knuckle. BRIEF DESCRIPTION OF THE DRAWINGS

[0047] Figure 1 is a schematic flow chart of the method of the present invention;

[0048] Figure 2 is a schematic flow chart of the time synchronization optimization algorithm of the present invention;

[0049] Figure 3 is a schematic flow chart of the multi-angle feature fusion algorithm of the present invention;

[0050] Figure 4 is a schematic three-dimensional structure diagram (one) of the multi-angles of the steering knuckle of the present invention;

[0051] Figure 5 is a schematic three-dimensional structure diagram (two) of the multi-angles of the steering knuckle of the present invention;

[0052] Figure 6 is a schematic three-dimensional structure diagram (three) of the multi-angles of the steering knuckle of the present invention;

[0053] Figure 7 is a schematic diagram of the production process of the steering knuckle of the present invention.

[0054] Description of the reference numerals in the drawings: 1 - steering knuckle to be detected, 2 - detection robotic arm, 3 - laser scanner, 4 - area array industrial camera, 5 - detected steering knuckle. DETAILED DESCRIPTION OF THE INVENTION

[0055] The following describes exemplary embodiments of the present disclosure with reference to the accompanying drawings. Various details of the embodiments of the present disclosure are included to facilitate understanding, and they should be considered merely exemplary. Therefore, those of ordinary skill in the art should recognize that various changes and modifications can be made to the embodiments described herein without departing from the scope and spirit of the present disclosure. Similarly, descriptions of well-known functions and structures are omitted in the following description for clarity and conciseness.

[0056] Embodiment 1: As Figure 1 or Figure 4 or Figure 5 or Figure 6 shown, a multi-angle detection method applied to a steering knuckle, the method comprising:

[0057] U1. On the production line of the steering knuckle, based on a planar array industrial camera, multi-angle image data information of the steering knuckle is obtained in real time, and based on a laser scanner, multi-angle point cloud data information of the steering knuckle is obtained in real time, and preprocessing is performed to obtain preprocessed multi-angle image data information and point cloud data information of the steering knuckle;

[0058] U2. Based on the preprocessed multi-angle image data information and point cloud data information of the steering knuckle, a time synchronization optimization algorithm is used to optimize the image and point cloud data information of the steering knuckle to obtain optimized multi-angle image data information and optimized multi-angle point cloud data information of the steering knuckle;

[0059] U3. Based on the optimized multi-angle image data information and optimized multi-angle point cloud data information of the steering knuckle, a multi-angle feature fusion algorithm is used to fuse the image and point cloud data information of the steering knuckle, and the fused multi-angle data information of the steering knuckle is output;

[0060] U4. Based on the fused multi-angle data information of the steering knuckle, a multi-angle matching algorithm of the steering knuckle is used to match with a preset standard model of the steering knuckle to obtain multi-angle matching degree data information of the steering knuckle;

[0061] U5. Based on the multi-angle matching degree data information of the steering knuckle, an evaluation function M of the steering knuckle is constructed to evaluate the matching degree of the steering knuckle, and multi-angle detection data information of the steering knuckle is obtained.

[0062] In this embodiment, in step U1, the preprocessing includes:

[0063] U11. Denoising, image enhancement, and image registration processing are performed on the multi-angle image data information of the steering knuckle to obtain preprocessed multi-angle image data information of the steering knuckle;

[0064] U12. Remove the outlier points, filter, and voxelize the multi-angle point cloud data information of the steering knuckle to obtain the preprocessed multi-angle point cloud data information of the steering knuckle.

[0065] In this embodiment, as Figure 2 shown, in step U2, the optimization of the image and point cloud data information of the steering knuckle using the time synchronization optimization algorithm includes:

[0066] U21. Based on the preprocessed multi-angle image data information and point cloud data information of the steering knuckle, perform timestamp calibration to obtain the calibrated multi-angle image data information of the steering knuckle and the calibrated multi-angle point cloud data information of the steering knuckle;

[0067] U22. Based on the calibrated multi-angle image data information and calibrated multi-angle point cloud data information of the steering knuckle, for each angle of the steering knuckle, establish the image time synchronization optimization function G of the steering knuckle and the point cloud time synchronization optimization function H of the steering knuckle,

[0068] ,

[0069] ,

[0070] where T is the sampling period, t is different moments, ζ is a constant parameter, X t is the calibrated multi-angle image data information of the steering knuckle at time t, Y t is the calibrated multi-angle point cloud data information of the steering knuckle at time t, X t+ζ is the calibrated multi-angle image data information of the steering knuckle at time t + ζ, Y t+ζ is the calibrated multi-angle point cloud data information of the steering knuckle at time t + ζ, process the calibrated multi-angle image data information and point cloud data information of the steering knuckle to obtain the processed multi-angle image data information of the steering knuckle and the processed multi-angle point cloud data information of the steering knuckle;

[0071] U23. Based on the processed multi-angle image data information of the steering knuckle and the processed multi-angle point cloud data information of the steering knuckle, synchronize according to the time series to obtain the optimized multi-angle image data information of the steering knuckle and the optimized multi-angle point cloud data information of the steering knuckle.

[0072] In this embodiment, the synchronization according to the time series is to make the processed multi-angle image data information of the steering knuckle and the processed multi-angle point cloud data information of the steering knuckle correspond one by one according to the time points, and eliminate the non-corresponding processed multi-angle image data information or processed multi-angle point cloud data information of the steering knuckle.

[0073] Embodiment 2: On the basis of a multi-angle detection method for a steering knuckle in Embodiment 1, the present invention will be further described and explained below.

[0074] As Figure 1 or Figure 4 or Figure 5 or Figure 6 shown, a multi-angle detection method for a steering knuckle, the method comprising:

[0075] U1. On the production line of the steering knuckle, based on a surface array industrial camera, multi-angle image data information of the steering knuckle is obtained in real time, and based on a laser scanner, multi-angle point cloud data information of the steering knuckle is obtained in real time, and preprocessing is performed to obtain preprocessed multi-angle image data information and point cloud data information of the steering knuckle;

[0076] U2. Based on the preprocessed multi-angle image data information and point cloud data information of the steering knuckle, a time synchronization optimization algorithm is used to optimize the image and point cloud data information of the steering knuckle, and optimized multi-angle image data information and optimized multi-angle point cloud data information of the steering knuckle are obtained;

[0077] U3. Based on the optimized multi-angle image data information and optimized multi-angle point cloud data information of the steering knuckle, a multi-angle feature fusion algorithm is used to fuse the image and point cloud data information of the steering knuckle, and fused multi-angle data information of the steering knuckle is output;

[0078] U4. Based on the fused multi-angle data information of the steering knuckle, a multi-angle matching algorithm of the steering knuckle is used to match with a preset standard model of the steering knuckle, and multi-angle matching degree data information of the steering knuckle is obtained;

[0079] U5. Based on the multi-angle matching degree data information of the steering knuckle, an evaluation function M of the steering knuckle is constructed to evaluate the matching degree of the steering knuckle, and multi-angle detection data information of the steering knuckle is obtained.

[0080] In this embodiment, as Figure 3 shown, in step U3, the use of the multi-angle feature fusion algorithm to fuse the image and point cloud data information of the steering knuckle includes:

[0081] U31. Input the optimized multi-angle image data information of the steering knuckle into a trained convolutional neural network model for feature point extraction, and multi-angle image feature point data information of the steering knuckle is obtained;

[0082] U32. Based on the multi-angle image feature point data information of the steering knuckle and the optimized multi-angle point cloud data information of the steering knuckle, a feature fusion function M is established,

[0083] ,

[0084] Among them, A is the multi - angle image feature point data information of the knuckle, B is the multi - angle point cloud data information of the optimized knuckle, and α and β are the feature fusion parameter matrices;

[0085] U33. Based on the feature fusion function M, obtain the multi - angle data information of the fused knuckle.

[0086] In this embodiment, in step U4, the adoption of the multi - angle matching algorithm of the knuckle to match with the preset standard model of the knuckle includes:

[0087] U41. Based on the multi - angle data information of the fused knuckle and the preset standard model of the knuckle, construct the matching matrix of the knuckle to obtain the multi - angle matching matrix data information of the fused knuckle and the matching matrix data information of the standard model;

[0088] U42. Based on the multi - angle matching matrix data information of the fused knuckle and the matching matrix data information of the standard model, establish the multi - angle matching function N of the knuckle,

[0089] ,

[0090] Among them, C is the multi - angle matching matrix data information of the fused knuckle, D is the matching matrix data information of the standard model, γ and δ are the transformation matrices, and µ is the matching constant parameter;

[0091] U43. Based on the multi - angle matching function N of the knuckle, obtain the multi - angle matching degree data information of the knuckle.

[0092] In this embodiment, the constraint condition of the matching constant parameter µ is:

[0093] .

[0094] In this embodiment, in step U5, construct the evaluation function M of the knuckle,

[0095] ,

[0096] Among them, n is the sample size, θ ij is the j - th matching degree data information of the knuckle at the i - th angle, and σ ij is the weight coefficient.

[0097] In this embodiment, the present invention provides a multi - angle detection system applied to a knuckle, including a computer device, which is programmed or configured to execute the steps of any one of the multi - angle detection methods applied to the knuckle.

[0098] As Figure 7 shown, the steering knuckle 1 to be detected is conveyed by a conveyor belt to the detection robotic arm 2. The detection robotic arm 2 grabs the steering knuckle to be detected, collects data through at least six area array industrial cameras 4 and a laser scanner 3, and detects the steering knuckle 1 to be detected. If the steering knuckle 1 to be detected meets the detection standard, the detection robotic arm 2 transfers the detected steering knuckle 5 to another conveyor belt for conveyance.

[0099] In this embodiment, the present invention provides a computer-readable storage medium, on which a computer program programmed or configured to execute the multi-angle detection method for a steering knuckle described in any one of the above is stored.

[0100] Any reference to a memory, storage, database, or other medium used in the embodiments provided in the present application may include non-volatile and / or volatile memory. Non-volatile memory may include read-only memory (ROM), programmable ROM (PROM), electrically programmable ROM (EPROM), electrically erasable programmable ROM (EEPROM), or flash memory. Volatile memory may include random access memory (RAM) or external cache memory. By way of illustration and not limitation, RAM is available in various forms, such as static RAM (SRAM), dynamic RAM (DRAM), synchronous DRAM (SDRAM), double data rate SDRAM (DDR SDRAM), enhanced SDRAM (ESDRAM), synchronous link DRAM (SLDRAM), Rambus direct RAM (RDRAM), direct memory bus dynamic RAM (DRDRAM), and Rambus dynamic RAM (RDRAM), etc.

[0101] In summary, the present invention not only solves the problem of low accuracy of the steering knuckle, but also the entire detection process is convenient and fast, and can further improve the production efficiency of the steering knuckle.

[0102] The above specific embodiments do not constitute a limitation on the protection scope of the present disclosure. Those skilled in the art should understand that various modifications, combinations, sub-combinations, and substitutions can be made according to design requirements and other factors. Any modifications, equivalent substitutions, and improvements made within the spirit and principle of the present disclosure shall be included within the protection scope of the present disclosure.

Claims

1. A multi-angle detection method applied to a steering knuckle, characterized in that, The method includes: U1. On the production line of the steering knuckle, based on the area array industrial camera, the multi-angle image data information of the steering knuckle is obtained in real time, and based on the laser scanner, the multi-angle point cloud data information of the steering knuckle is obtained in real time, and preprocessing is performed to obtain the preprocessed multi-angle image data information and point cloud data information of the steering knuckle; U2. Based on the preprocessed multi-angle image data information and point cloud data information of the steering knuckle, the time synchronization optimization algorithm is used to optimize the image and point cloud data information of the steering knuckle, and the optimized multi-angle image data information of the steering knuckle and the optimized multi-angle point cloud data information of the steering knuckle are obtained; U3. Based on the optimized multi-angle image data information of the steering knuckle and the optimized multi-angle point cloud data information of the steering knuckle, the multi-angle feature fusion algorithm is used to fuse the image and point cloud data information of the steering knuckle, and the fused multi-angle data information of the steering knuckle is output; U4. Based on the fused multi-angle data information of the steering knuckle, the multi-angle matching algorithm of the steering knuckle is used to match with the preset standard model of the steering knuckle, and the multi-angle matching degree data information of the steering knuckle is obtained; U5. Based on the multi-angle matching degree data information of the steering knuckle, an evaluation function M of the steering knuckle is constructed to evaluate the matching degree of the steering knuckle, and the multi-angle detection data information of the steering knuckle is obtained; In step U2, the use of the time synchronization optimization algorithm to optimize the image and point cloud data information of the steering knuckle includes: U21. Based on the preprocessed multi-angle image data information and point cloud data information of the steering knuckle, timestamp calibration is performed to obtain the calibrated multi-angle image data information of the steering knuckle and the calibrated multi-angle point cloud data information of the steering knuckle; U22. Based on the calibrated multi-angle image data information of the steering knuckle and the calibrated multi-angle point cloud data information of the steering knuckle, for each angle of the steering knuckle, an image time synchronization optimization function G of the steering knuckle and a point cloud time synchronization optimization function H of the steering knuckle are established; ， ， Among them, T is the sampling period, t is different moments, ζ is a constant parameter, X t is the multi-angle image data information of the calibrated steering knuckle at time t, Y t is the multi-angle point cloud data information of the calibrated steering knuckle at time t, X t+ζ is the multi-angle image data information of the calibrated steering knuckle at time t + ζ, Y t+ζ is the multi-angle point cloud data information of the calibrated steering knuckle at time t + ζ. Process the multi-angle image data information and the point cloud data information of the calibrated steering knuckle to obtain the processed multi-angle image data information of the steering knuckle and the processed multi-angle point cloud data information of the steering knuckle; U23. Based on the processed multi-angle image data information of the steering knuckle and the processed multi-angle point cloud data information of the steering knuckle, synchronization is performed according to the time sequence to obtain the optimized multi-angle image data information of the steering knuckle and the optimized multi-angle point cloud data information of the steering knuckle; In step U3, the use of the multi-angle feature fusion algorithm to fuse the image and point cloud data information of the steering knuckle includes: U31. Input the optimized multi-angle image data information of the steering knuckle into the trained convolutional neural network model for feature point extraction to obtain the multi-angle image feature point data information of the steering knuckle; U32. Based on the multi-angle image feature point data information of the steering knuckle and the optimized multi-angle point cloud data information of the steering knuckle, a feature fusion function M is established; ， where A is the multi-angle image feature point data information of the steering knuckle, B is the optimized multi-angle point cloud data information of the steering knuckle, and α and β are feature fusion parameter matrices; U33. Based on the feature fusion function M, the fused multi-angle data information of the steering knuckle is obtained; In step U4, the multi-angle matching algorithm of the steering knuckle is matched with the preset standard model of the steering knuckle, which includes: U41. Based on the multi-angle data information of the fused steering knuckle and the preset standard model of the steering knuckle, construct a matching matrix of the steering knuckle to obtain the multi-angle matching matrix data information of the fused steering knuckle and the matching matrix data information of the standard model; U42. Based on the multi-angle matching matrix data information of the fused steering knuckle and the matching matrix data information of the standard model, establish a multi-angle matching function N of the steering knuckle, ， where C is the multi-angle matching matrix data information of the fused steering knuckle, D is the matching matrix data information of the standard model, γ and δ are transformation matrices, and µ is a matching constant parameter; U43. Based on the multi-angle matching function N of the steering knuckle, obtain the multi-angle matching degree data information of the steering knuckle.

2. The multi-angle detection method applied to a steering knuckle according to claim 1, characterized in that, In step U1, the preprocessing includes: U11. Perform denoising, image enhancement, and image registration processing on the multi-angle image data information of the steering knuckle to obtain the preprocessed multi-angle image data information of the steering knuckle; U12. Perform outlier removal, filtering, and voxelization processing on the multi-angle point cloud data information of the steering knuckle to obtain the preprocessed multi-angle point cloud data information of the steering knuckle.

3. The multi-angle detection method applied to a steering knuckle according to claim 1, characterized in that: The synchronization according to the time series is to make the processed multi-angle image data information of the steering knuckle and the processed multi-angle point cloud data information of the steering knuckle correspond one by one according to the time points, and eliminate the multi-angle image data information or multi-angle point cloud data information of the steering knuckle that cannot be corresponded.

4. The multi-angle detection method applied to a steering knuckle according to claim 1, characterized in that, The constraint condition of the matching constant parameter µ is: 。 5. The multi-angle detection method applied to a steering knuckle according to claim 1, characterized in that, In step U5, construct an evaluation function M of the steering knuckle, ， where n is the sample size, and θ ij is the j-th matching degree data information of the steering knuckle at the i-th angle, and σ ij is the weight coefficient.

6. An angular multi-detection system applied to a steering knuckle, comprising a computer device, characterized in that, The computer device is programmed or configured to execute the steps of the multi-angle detection method for the steering knuckle according to any one of claims 1 to 5.

7. A computer-readable storage medium, characterized in that, The computer-readable storage medium stores a computer program programmed or configured to execute the multi-angle detection method for the steering knuckle according to any one of claims 1 to 5.

Images