Non-contact automobile part testing fixture system based on blue light scanning technology and operation method of non-contact automobile part testing fixture system

Abstract

The invention discloses a non-contact automobile part checking fixture system based on a blue light scanning technology and an operation method thereof, the system comprises a hardware part and a software part, the hardware part is responsible for data acquisition and transmission, the software part is responsible for data processing and analysis work, the hardware part comprises a blue light scanner and an equipment transmission and processing unit, and the equipment transmission and processing unit is responsible for the blue light scanning. The blue light scanner obtains three-dimensional information of the surface of a part by emitting blue light and receiving reflected light, the equipment transmission and processing unit comprises an equipment transmission unit and a data processing unit, and the data processing unit is provided with a powerful central processing unit (CPU), a graphics processing unit (GPU) and a large-capacity memory. The software part comprises a scanning data acquisition module, a data preprocessing module, a three-bit comparison analysis module, a data management and tracing module and an equipment calibration and maintenance module. According to the invention, the abrasion and maintenance cost of equipment are reduced, the material waste and rework cost caused by waste products are reduced, the photon flux density is improved by 30%, and a signal missing area is reduced.

Description

Technical Field

[0001] This invention relates to the field of automotive parts inspection technology, specifically to a non-contact automotive parts inspection tool system based on blue light scanning technology and its operation method. Background Technology

[0002] In the automotive manufacturing industry, high-precision inspection of parts and fixtures plays a decisive role in ensuring the quality and performance of automobiles. Traditional inspection methods are gradually revealing many limitations when faced with the high-precision requirements of modern automobile manufacturing.

[0003] Manual inspection relies heavily on the experience and skill level of the inspectors. Differences in operating habits and judgment standards among personnel make it difficult to guarantee the consistency and accuracy of inspection results. Furthermore, manual inspection is inefficient and struggles to meet the rapidly growing production demands in the context of large-scale automotive parts manufacturing. Using conventional measuring tools such as calipers and micrometers not only limits the measurement range but also makes it difficult to perform comprehensive and accurate measurements on complex-shaped parts. For example, the internal structure of an automotive engine block is complex, containing numerous irregular curved surfaces and holes. Conventional measuring tools simply cannot reach certain critical areas, and the measurement accuracy is insufficient to meet the requirements of automotive manufacturing.

[0004] While coordinate measuring machines (CMMs) offer improved accuracy, their measurement speed is relatively slow. Measuring a moderately complex automotive part often takes several hours or even longer, severely impacting production efficiency. Furthermore, CMMs typically only allow for sampling inspection of parts, failing to achieve full-size, comprehensive inspection. This may result in the omission of defective products, rendering traditional inspection methods inadequate to meet this evolving need.

[0005] Specifically, existing technologies for measuring the special surfaces and structures of automotive parts have the following problems: (1) Commonly used automotive parts include highly reflective metals (such as aluminum alloy wheels with reflectivity > 80%) and strong light-absorbing materials (such as carbon fiber interior parts with reflectivity < 5%), which can lead to serious data distortion: the former causes overexposure of light spots, and the latter causes signal loss. Both can form a measurement blind zone with an area of ​​more than 10%. (2) For complex structures such as deep holes (e.g., φ5mm deep holes in engine cylinder blocks) and narrow slits (e.g., gear meshing clearance in gearboxes), traditional scanning heads cannot achieve effective scanning due to physical size limitations (minimum diameter 50mm) and viewing angle obstruction, with blind spots accounting for more than 30%; (3) The accuracy of capturing small features such as threads and chamfers is insufficient, and the feature boundary offset is often ≥0.3mm. (4) Existing traditional contact inspection technologies (such as coordinate measuring machines, calipers, etc.) rely on the direct contact between the probe and the surface of the part, which has two major limitations: First, it is easy to cause scratches and indentations on easily deformable and precision-surfaced parts (such as aluminum alloy body frames and plastic interior parts), resulting in an increased scrap rate (the industry average scrap rate is about 2%-5%); Second, it can only measure a single point at a time, and for complex curved surface parts (such as engine cylinder blocks and headlight covers), the probe position needs to be adjusted multiple times, resulting in a long inspection cycle. (5) Existing traditional optical inspection technology has two major shortcomings: First, the accuracy is easily affected by ambient light (such as changes in natural light in the workshop and equipment lighting, which can cause the measurement error to increase to more than ±0.1mm), and the scanning effect on dark and highly reflective parts (such as black plastic bumpers and stainless steel exhaust pipes) is poor, which can easily lead to data loss; Second, data processing depends on external computers, and data needs to be manually imported and analyzed offline. The interval from scanning to generating the inspection report is as long as 1-2 hours, which cannot meet the needs of real-time quality control. (6) The current inspection of some automotive parts (such as large body panels and complex pipelines) still relies on "manual hand-held inspection tools + visual judgment", which has two major problems: First, manual operation is easily affected by subjective factors (such as fatigue of inspectors and differences in experience), resulting in measurement errors (such as gap measurement error can reach ±0.2mm), and appearance defects (such as minor scratches) are easily missed; Second, manual inspection requires professional training (the training period is about 3-6 months), and each inspection requires 1-2 people to cooperate, which results in high labor costs and makes it difficult to adapt to the "large-scale and standardized" production needs of the automotive industry.

[0006] Therefore, there is an urgent need for a more efficient, accurate, and comprehensive inspection and scanning technology, which has brought new solutions to the inspection of automotive parts. Summary of the Invention

[0007] This invention provides a non-contact automotive parts inspection system and its operation method based on blue light scanning technology, which solves the problem of easily scratched surfaces of some vulnerable parts, such as plastic parts in automotive interiors. The non-contact nature of blue light scanning avoids damage to the surface of parts caused by physical contact, ensuring the integrity and original performance of the parts, and providing a reliable data foundation for subsequent quality inspection and analysis.

[0008] This application is achieved through the following technical solution: A non-contact automotive parts inspection system based on blue light scanning technology includes hardware and software components. The hardware component is responsible for data acquisition and transmission, while the software component handles data processing and analysis. The hardware component includes a blue light scanner and a device transmission and processing unit. The blue light scanner acquires three-dimensional information of the part surface by emitting blue light and receiving reflected light. The blue light scanner includes a blue light emitting device and a high-resolution industrial camera. The blue light emitting device emits blue light of specific wavelengths and intensities to form a structured light pattern projected onto the part surface. The high-resolution industrial camera simultaneously captures the blue light pattern reflected from the part surface from different angles. The device transmission and processing unit includes a device transmission unit and a data processing unit. The device transmission unit is responsible for quickly and accurately transmitting the point cloud data acquired by the blue light scanner to the data processing unit. The data processing unit is a high-performance computer equipped with a powerful CPU, GPU, and large-capacity memory. The CPU handles data logic... The GPU is mainly used to accelerate graphics processing and 3D model calculations, while large-capacity memory is used to store and quickly retrieve large amounts of scan data and intermediate calculation results. The software part includes a scan data acquisition module, a data preprocessing module, a 3D comparison and analysis module, a data management and traceability module, and an equipment calibration and maintenance module. The scan data acquisition module is used to control the working parameters and scanning process of the blue light scanner, realizing precise operation of the hardware device. It interacts with the hardware driver of the blue light scanner to control the hardware device, send control commands and receive status information from the hardware to ensure the smooth progress of the scanning process. The data analysis and detection software is responsible for in-depth analysis and processing of the received scan data to realize quality inspection of automotive parts. The data preprocessing module first accurately compares the scanned point cloud data with the pre-imported model, and then calculates the deviation between the actual size and the design size of the part, including shape deviation and size deviation, through a specific algorithm.

[0009] In a preferred embodiment, the device transmission unit adopts a high-speed data transmission interface, such as Ethernet or USB 3.0, to meet the real-time transmission requirements of large data volumes.

[0010] In a preferred embodiment, the point cloud data collected by the blue light scanner is transmitted to a computer in real time via a data cable. The data processing software in the computer performs preprocessing operations such as noise reduction, filtering, and alignment on the data, and then performs comparative analysis with the model to finally generate a detection report.

[0011] In a preferred embodiment, the scanning control software sets the scanning frequency. Based on the complexity of the part and the required detection accuracy, it selects an appropriate scanning frequency to balance scanning efficiency and data quality. By adjusting the exposure time, it ensures that the high-resolution industrial camera can accurately capture the blue light signal reflected from the surface of the part, avoiding overexposure or underexposure. When scanning automotive interior parts, which have diverse colors and materials, it is necessary to finely adjust the exposure time according to the surface characteristics of the part. The scanning control software has a scanning path planning function, which automatically generates the optimal scanning path based on the shape and size of the part, ensuring that every area of ​​the part surface can be scanned completely and accurately.

[0012] In a preferred embodiment, the data analysis and testing software for inspecting automotive body panels carefully compares the scanned data and models of the panels to accurately identify potential issues such as localized deformation and dimensional deviations. Based on the comparison results, the software automatically generates detailed inspection reports. These reports include various inspection data of the parts, deviation analysis charts, and conclusions on whether the parts are qualified. These inspection reports provide important information for automotive manufacturers' quality control and production decisions, helping them to promptly identify and resolve parts quality issues, thereby improving product quality and production efficiency.

[0013] In a preferred embodiment, the blue light scanner is used for structured light projection and image acquisition. Specifically, the blue light scanner uses a specific optical device to project a carefully designed grating pattern onto the surface of an automotive part in the form of blue light. These grating patterns have periodic stripe structures, such as sinusoidal stripes or Gray code stripes. Taking sinusoidal stripes as an example, the brightness of the stripes changes according to a sine function. When blue light is projected onto the surface of the part, the originally regular grating pattern will be deformed due to the undulations of the part's surface. Two high-resolution industrial cameras simultaneously acquire these deformed stripe images from different angles. The position and angle of the high-resolution industrial cameras are precisely calibrated to ensure that the deformation information of each area of ​​the part's surface can be accurately captured. When scanning automotive engine blades, blue light is projected onto the blade surface. The complex curvature of the blade causes the grating pattern to produce unique deformations. Two high-resolution industrial cameras quickly capture these deformed stripes from different perspectives, providing raw image data for subsequent three-dimensional coordinate calculations. In a preferred embodiment, a fixed geometric relationship exists between the high-resolution industrial camera and the projector, including their baseline distance and relative angle. When blue light is projected onto the surface of the part and reflected, the pixels in the deformed fringe image received by the high-resolution industrial camera correspond to the original fringe projected by the projector. By analyzing these correspondences and using trigonometric functions and geometric relationships, the three-dimensional coordinates of each point on the surface of the part can be calculated. Assuming the optical center of the high-resolution industrial camera is O, the optical center of the projector is P, a point on the surface of the part is M, the baseline distance between the high-resolution industrial camera and the projector is b, and the imaging point of point M on the image plane captured by the high-resolution industrial camera is m, by analyzing the deformation of the fringe in the image, the angle information θ corresponding to point m can be obtained. According to the trigonometric relationship, in triangle OMP, given OP = b and θ, the distance z from point M to the plane where the high-resolution industrial camera and the projector are located can be calculated. Combining the imaging model of the high-resolution industrial camera and other geometric parameters, the x and y coordinates of point M in three-dimensional space can be further calculated, thereby achieving accurate measurement of the three-dimensional coordinates of each point on the surface of the part.

[0014] In a preferred embodiment, the image data acquired by the blue light scanner is first converted into point cloud data through a specific algorithm. During this process, the deformed stripes in the image are decoded and analyzed to extract the three-dimensional coordinate information corresponding to each pixel, thereby generating point cloud data containing a large number of discrete points. Since the actual scanning process may be affected by environmental noise and equipment errors, there are often some noise points and outliers in the point cloud data. Therefore, it is necessary to perform noise reduction processing. Common noise reduction methods include Gaussian filtering, which smooths the point cloud data and removes noise by weighted averaging of each point and its neighborhood. Bilateral filtering can also be used, which not only considers the spatial distance of points but also the differences in point attributes, and can effectively remove noise while preserving the detailed features of the point cloud. In addition, filtering operations such as voxel filtering are also performed. This divides the point cloud space into small voxels and takes the average value of the points in each voxel as a representative point, thereby reducing the amount of point cloud data and improving the efficiency of subsequent processing. In a preferred embodiment, the data analysis and inspection software aligns and compares the preprocessed point cloud data with the original model of the automotive part. It employs the Iterative Closest Point (ICP) algorithm, continuously iterating to achieve optimal spatial matching between the point cloud data and the original model. During alignment, the distance from each point in the point cloud data to the surface of the original model is calculated. Based on this distance information, the deviation between the actual part and the design model is determined. A specific algorithm calculates the deviation values ​​between the actual shape and the designed shape of the part, such as dimensional and shape deviations. For automotive wheel hub inspection, comparative analysis can accurately determine the deviations between the wheel hub's diameter, spoke shape, and position parameters and the design values. Based on the set tolerance range, the software determines whether the part is qualified and generates a detailed inspection report. The inspection report displays the deviation distribution in an intuitive chart format, providing accurate quality inspection results for automotive manufacturers.

[0015] The operation method of the non-contact automotive parts inspection system based on blue light scanning technology includes the following steps: S1. Part Positioning and Fixing: Before scanning, the operator will select a suitable measuring bracket based on the shape, size and structural characteristics of the automotive parts. For parts with complex shapes, such as irregular sheet metal parts of the car body, a customized flexible fixture may be used. Adjustable positioning blocks and vacuum adsorption devices are used to achieve stable fixation of the parts. During the fixation process, the operator will carefully check the fit between the parts and the positioning device to ensure that the parts will not be displaced in any direction, so as to ensure the stability and accuracy of the part position during subsequent scanning. S2. Scanner Start-up and Scanning: After the parts are fixed in place, the operator starts the blue light scanner through the scanning control software. During startup, the software automatically initializes and checks various scanner parameters to ensure that the parameter settings meet the requirements of this scanning task. The scanning frequency can be adjusted according to the complexity of the parts. For parts with simple shapes and few surface features, a higher scanning frequency can be selected to improve scanning efficiency; for parts with complex shapes and rich surface details, a lower scanning frequency is selected to ensure that the subtle features of the part's surface can be captured. The exposure time is also optimized according to the surface material and color of the parts. For strong metal parts, shorten the exposure time appropriately to avoid overexposure; for plastic parts with good light absorption, extend the exposure time appropriately to ensure that the camera can accurately capture the blue light signal reflected from the surface of the part. During the scanning process, the blue light scanner scans the part from multiple angles and in all directions according to the preset scanning path. The scanning path is planned based on the three-dimensional model of the part to ensure that the entire surface of the part is covered and to avoid scanning blind spots. The scanner emits blue light from different angles and receives the reflected light, and simultaneously captures the blue light pattern reflected from the surface of the part. Using the principle of triangulation, the three-dimensional coordinates of each point on the surface of the part are quickly calculated to obtain complete surface data. S3. Data transmission and reception: After scanning, the large amount of point cloud data collected by the blue light scanner will be transmitted to the computer quickly through a high-speed data transmission line and a report generation system. During the transmission process, the data will be verified and error corrected to ensure the integrity and accuracy of the data. The data analysis and detection software in the computer will receive the data in real time and store it in a designated memory area or hard disk space. S4. Analysis and Report Generation: The data analysis and inspection software first preprocesses the received data, including noise reduction and filtering, to remove interference data caused by environmental noise and equipment errors, thus improving data quality. Next, the software accurately compares the preprocessed point cloud data with the pre-imported original model of the automotive part. Using a specific algorithm, it calculates the deviation between the actual size and design size of the part, including shape and dimensional deviations. For automotive doors, the software analyzes in detail the deviations in the door's contour, edges, mounting holes, etc. Based on the comparison results, the software automatically identifies out-of-tolerance areas and marks them intuitively, such as using different colored blocks to represent different degrees of deviation. Finally, the software generates an inspection report containing the inspection results and deviation analysis. The report not only includes various inspection data and deviation analysis charts but also clearly states whether the part is qualified. These reports can be exported in common formats such as PDF and Excel, making it convenient for quality control personnel in automotive manufacturing companies to view and analyze, providing strong support for the company's production decisions.

[0016] In a preferred embodiment, the specific algorithm used in step S3 includes: Algorithm 1: Noise Removal Algorithm: Eliminates interfering points and preserves true features. The statistical filtering algorithm Statistical Outlier Removal (SOR) calculates the average distance k of the k nearest neighbors for each point in the point cloud. The k value is usually set to 20-50 and adjusted according to the point cloud density. Then, it calculates the standard deviation based on the average distance of all points and identifies and deletes points whose distance is greater than the average distance + 2-3 times the standard deviation as noise points. Algorithm 2: Data Resampling Algorithm: Optimizes point cloud density and improves computational efficiency. The Voxel Grid Downsampling algorithm divides the three-dimensional space into uniformly sized "voxel grids". The voxel size is set according to the detection accuracy. For example, when the detection accuracy is ±0.1mm, the voxel size is set to 0.05-0.1mm. For all point cloud points in each voxel, the average value of its three-dimensional coordinates is calculated. This average value is used to represent the points of the entire voxel, and other points in the voxel are deleted. Algorithm 3: Point Cloud Segmentation Algorithm: Extract the target part and eliminate background interference. First, select seed points, which are usually feature points on the surface of the target part, such as the center of a hole or the edge of the part. Set growth conditions, such as the angle between the normal vector of the adjacent point and the seed point is less than 10° and the distance is less than 2mm. Then, starting from the seed point, gradually include the neighboring points that meet the growth conditions into the target area until no more growth is possible. Finally, the point cloud of the target part is obtained, and the background area is excluded.

[0017] In a preferred embodiment, step S4, which involves using a specific algorithm, specifically includes: Algorithm 1: Best Fit Alignment Algorithm. This algorithm extracts key feature points from the actual point cloud of the part, such as hole centers, part edge vertices, and surface feature points. Simultaneously, it extracts corresponding theoretical feature points from the digital model. The algorithm obtains these features through the original digital model feature extraction function and constructs an error function. The objective is to minimize the sum of squared distances between actual and theoretical feature points. The algorithm iteratively optimizes the transformation matrix using methods such as least squares and singular value decomposition. The optimal transformation matrix includes translations X / Y / Z and rotation angles α / β / γ. Algorithm 2: Point-to-point deviation algorithm. For each point in the actual point cloud, find the nearest theoretical point in the digital model. Through spatial distance calculation, select the digital model vertex or sampling point with the smallest distance and calculate the Euclidean distance between the two points. This distance is the point-to-point deviation of the point. If the deviation is positive, it means that the actual point is outside the theoretical point in the digital model; if it is negative, it means that the actual point is inside. Algorithm 3: Deviation Statistical Analysis Algorithm, which performs statistical calculations on all local deviation data to generate key statistical indicators, including: 1 / Mean Deviation: The arithmetic mean of all deviation values, reflecting the average offset direction of the part from the digital model. For example, a mean of +0.02mm indicates that the part is slightly larger than the design size. 2 / Standard Deviation: Reflects the degree of dispersion of deviation values. The smaller the standard deviation, the better the consistency of part dimensions. For example, the standard deviation of the piston of an automobile engine is usually required to be <0.008mm. 3 / Max Positive Deviation: The maximum value among all positive deviations, reflecting the most prominent area of ​​the part, such as a local bulge in a body panel. 4 / Max Negative Deviation: The absolute value of the minimum of all negative deviations is the largest, reflecting the most concave area of ​​the part, such as shrinkage marks in injection molded parts; 5 / Out-of-Tolerance Rate: The proportion of out-of-tolerance points to the total number of points. The absolute value of the deviation of out-of-tolerance points is greater than the upper limit of tolerance. The lower the out-of-tolerance rate, the higher the pass rate of the parts. For example, the out-of-tolerance rate requirement for key automotive parts is <0.5%.

[0018] Technical Implementation Principle: This invention is a non-contact automotive parts inspection system based on blue light scanning technology. It utilizes blue light projection and camera acquisition to achieve non-contact measurement. Its principle is based on optical triangulation. The system projects a specific pattern of blue light structured light onto the surface of the automotive parts. The contour of the part's surface causes the blue light stripes to deform. Simultaneously, a high-resolution industrial camera synchronously acquires images of these deformed blue light stripes from different angles. The core technology of the blue light scanning automotive parts inspection system focuses on the optimized application of a unique optical triangulation method, addressing the measurement principle. At the algorithm level, fast filtering algorithms and high-precision 3D reconstruction algorithms in image data processing are the core of achieving high-speed data processing and high-precision measurement. Based on the principle of optical triangulation, by calculating the triangular relationship between the camera and various points on the part's surface, the 3D coordinate information of each point on the part's surface can be accurately calculated, thus achieving non-contact measurement of the part's surface morphology. This non-contact measurement method has unique advantages for the inspection of complex and fragile parts. In automobile manufacturing, many parts have complex free-form surfaces, such as the intake manifold of a car engine. Its internal air passage structure is complex, and traditional contact measurement methods struggle to reach narrow or deep openings, potentially causing part deformation due to contact force and affecting measurement accuracy. However, the blue light scanning system can easily handle such complex structures, acquiring surface data from all angles without blind spots. It possesses high-speed data acquisition and processing capabilities. During the data acquisition phase, high-performance blue light projection equipment and a high-speed camera can acquire massive amounts of surface data in a short time. For example, for a medium-sized automotive part, this invention's non-contact automotive part inspection system based on blue light scanning technology can complete a full scan within seconds, collecting data from millions of measurement points. This data is rapidly transmitted to a computer in point cloud form. In the data processing stage, this invention's non-contact automotive part inspection system based on blue light scanning technology is equipped with an advanced hardware computing platform and optimized data processing algorithms. The inspection fixture system of this invention achieves high-precision measurement through various technical means. At the hardware level, the high-resolution industrial camera equipped with the system is one of the key factors. High-resolution industrial cameras have more pixels, enabling them to capture more subtle changes in blue light stripes, thereby improving measurement accuracy. For example, some high-end blue light scanning systems use industrial cameras with resolutions reaching tens of millions of pixels. Combined with precision optical lenses, they can clearly distinguish minute features and details on the surface of parts. Utilizing a high-precision 3D reconstruction algorithm, the 3D coordinates of each point on the surface of the part are accurately calculated based on the image data acquired by the camera, achieving sub-millimeter level or even higher precision measurements. The inspection fixture system of this invention achieves deep integration with the original digital model. The blue light scanning system and the original digital model are deeply integrated. In the actual inspection process, the original digital model of the automotive part is first imported into the inspection software as a standard model. Then, the system performs blue light scanning on the actual part to obtain the 3D point cloud data of the part.Inspection software precisely matches and aligns the scanned point cloud data with the original digital model, and uses specific algorithms to calculate the deviation between the two. For example, in the inspection of automotive body panels, after comparing the scanned point cloud data of the body panel with the original digital model, the software generates a color deviation map, visually displaying the deviations of various parts of the part from the design model. Green areas indicate deviations within the allowable range, while yellow and red areas indicate deviations exceeding the tolerance range. Technicians can quickly and accurately judge the manufacturing quality of parts based on this visualized information, analyze the location and extent of deformation, dimensional deviations, and other problems that occur during stamping, welding, and other processing, and then make targeted adjustments and optimizations to the production process.

[0019] Beneficial Effects: This invention, a non-contact automotive parts inspection system based on blue light scanning technology, utilizes a multi-core high-performance processor and large-capacity memory in terms of hardware, enabling rapid processing of large-scale point cloud data. Algorithmically, it employs parallel computing, fast filtering, and feature extraction techniques to achieve real-time analysis of the collected data. Compared to traditional inspection methods, such as coordinate measuring machines (CMMs), which require point-by-point measurement and are slow (often requiring tens of minutes or even hours to complete a full inspection of a part), the blue light scanning system can complete the same task in minutes, significantly improving inspection efficiency. The optimized software algorithm also plays a crucial role in high-precision measurement. Through advanced image processing algorithms, it can accurately identify and extract the feature information of blue light stripes, reducing the impact of noise and interference on the measurement results. Specifically, this is reflected in the following aspects: (a) Improved detection efficiency In traditional automotive parts inspection, coordinate measuring machines (CMMs) are commonly used inspection equipment. Their inspection process relies on the contact between the probe and the surface of the part to measure and acquire data point by point.

[0020] (ii) Improved detection accuracy High-precision testing is crucial for automobile manufacturing. Taking automobile wheel hubs as an example, their manufacturing precision directly affects the stability and safety of the vehicle. (iii) Reduced testing costs From the perspective of labor costs, traditional testing methods require a large number of professional testing personnel to operate, while the blue light scanning system has a high degree of automation, requiring only a small number of technicians to operate and monitor, which greatly reduces labor costs. The shortened testing time also reduces the cost of equipment use, reduces equipment wear and maintenance costs, and reduces the defect rate through high-precision testing, thus reducing material waste and rework costs caused by scrap.

[0021] (iv) Promoted innovation in automobile manufacturing technology Blue light scanning systems provide rich data support for automotive design and manufacturing process improvement. In the automotive design phase, blue light scanning of prototype parts acquires actual data, which is then compared and analyzed with the design model to identify potential problems in the design and optimize the design scheme. In terms of manufacturing process improvement, such as stamping and welding processes for automotive parts, blue light scanning systems are used to inspect parts before and after processing, analyze the impact of process parameters on the dimensional accuracy and shape of parts, thereby optimizing process parameters and improving the level of manufacturing process.

[0022] (v) The adoption of a multimodal scanning module overcomes the limitations of special measurements. Surface-adaptive scanning unit: For highly reflective surfaces, it integrates an HDR high dynamic range imaging module, reducing the exposure time from the conventional 25-35ms to 15-22ms, and with 3 levels of dynamic range extension, avoids signal overexposure; For surfaces with strong light absorption, pulsed blue light enhancement technology is used to increase photon flux density by 30%, while cross-scanning path optimization reduces signal loss areas. Miniaturized detection components: A miniature scanning probe with a diameter of 15mm was developed and mounted on a flexible robotic arm to achieve axial compensation scanning of deep holes with a diameter of less than 10mm. Depth correction was performed using the formula Δd' = Δd × (1+Ra / 50) (Ra is the surface roughness), so that the blind zone ratio of deep hole measurement is less than 1%. Feature enhancement algorithm: Introducing an AI-driven feature recognition engine to automatically locate and densify point clouds of minute features such as threads and chamfers (density increased to 200 points / mm). 2 This will control the feature boundary matching error to within 0.08mm. Attached Figure Description

[0023] Figure 1 This is a schematic diagram of the hardware data processing unit of the present invention.

[0024] Figure 2 This is a side view of the data processing unit in the hardware section of the present invention.

[0025] Figure 3 This is a schematic diagram of the overall structure of the software modules in this invention.

[0026] Figure 4 This is a schematic diagram of the signal transmission control structure of the software part in this invention.

[0027] Figure 5 This is a flowchart illustrating the operation of the inspection system of the present invention. Detailed Implementation

[0028] The embodiments of the present invention will be described in detail below with reference to the accompanying drawings: These embodiments are implemented based on the technical solution of the present invention, and provide detailed implementation methods and specific operation processes, but the protection scope of the present invention is not limited to the following embodiments.

[0029] like Figure 1 , 2 As shown, the non-contact automotive parts inspection system based on blue light scanning technology includes hardware and software components. The hardware component is responsible for data acquisition and transmission, and includes a blue light scanner and a device transmission and processing unit.

[0030] Blue light scanner: As the core hardware of the system, the blue light scanner acquires three-dimensional information of the part surface by emitting blue light and receiving reflected light. Its core components include a blue light emitting device and a high-resolution industrial camera. The blue light emitting device is responsible for emitting blue light of a specific wavelength and intensity to form a structured light pattern that is projected onto the part surface, while the high-resolution industrial camera simultaneously captures the blue light pattern reflected from the part surface from different angles. The camera typically has high pixel count and high frame rate characteristics, enabling it to capture subtle features and changes on the part surface. Working together, using the principle of triangulation, the two devices calculate the displacement and angular changes of the blue light pattern on the camera's imaging plane to accurately calculate the three-dimensional coordinates of each point on the part surface, thereby generating point cloud data of the part surface, providing a basis for subsequent inspection and analysis.

[0031] Blue light scanners utilize specialized optical devices to project meticulously designed grating patterns as blue light onto the surface of automotive parts. These grating patterns typically possess periodic stripe structures, such as sinusoidal stripes or Gray code stripes. Taking sinusoidal stripes as an example, the brightness of the stripes varies according to a sine function. When blue light is projected onto the part's surface, the originally regular grating pattern is deformed due to the surface's undulations. Two high-resolution industrial cameras simultaneously acquire these deformed stripe images from different angles. The camera positions and angles are precisely calibrated to ensure accurate capture of deformation information in various areas of the part's surface. When scanning automotive engine blades, blue light is projected onto the blade surface. The blade's complex curvature causes unique deformations in the grating pattern. Two industrial cameras rapidly capture these deformed stripes from different perspectives, providing raw image data for subsequent 3D coordinate calculations. Based on the principle of triangulation, a fixed geometric relationship exists between the camera and the projector, including their distance (baseline distance) and relative angles. After the blue light is projected onto the part's surface and reflected, the pixels in the deformed stripe image received by the camera correspond to the original stripes projected by the projector. By analyzing these correspondences and utilizing trigonometric functions and geometric relationships, the three-dimensional coordinates of each point on the part's surface can be calculated. Assume the camera's optical center is O, the projector's optical center is P, a point on the part's surface is M, and the baseline distance between the camera and the projector is b. The image point of point M on the image plane captured by the camera is m. By analyzing the deformation of the fringes in the image, the angle information θ corresponding to point m can be obtained. Based on trigonometric relationships, in triangle OMP, given OP = b and θ, the distance z from point M to the plane containing the camera and projector can be calculated. Combining this with the camera's imaging model and other geometric parameters, the x and y coordinates of point M in three-dimensional space can be further calculated, thus achieving accurate measurement of the three-dimensional coordinates of each point on the part's surface.

[0032] Device Transmission and Processing Unit: The device transmission unit is responsible for quickly and accurately transmitting the point cloud data acquired by the blue light scanner to the data processing unit. It typically uses high-speed data transmission interfaces such as Ethernet or USB 3.0 to meet the real-time transmission requirements of large data volumes. Figure 1 , 2As shown, the data processing unit is a high-performance computer equipped with a powerful CPU, GPU, and large-capacity memory. The CPU is responsible for the logical processing of data and the execution of analysis algorithms, while the GPU is mainly used to accelerate graphics processing and 3D model calculations. The memory is used to store and quickly retrieve large amounts of scan data and intermediate calculation results. In actual work, the point cloud data collected by the blue light scanner is transmitted to the computer in real time via a data cable. The data processing software in the computer performs preprocessing operations such as noise reduction, filtering, and alignment on the data, and then performs comparative analysis with the model to finally generate a detection report. The figure includes the host and the monitor, which are connected together by the top cabinet and the monitor connecting shaft 2. The monitor has a monitor bracket connection point 1 for fixing the monitor bracket, and the monitor has internal hierarchical fixing buckles 3. The top of the host has an internal interface module 3 and a control panel 11. The upper part of the side of the host has a cooling fan bracket 5 with a heat dissipation grille 12 installed on it. The lower part of the side of the host has a right-side panel heat dissipation grille support structure 10. The bottom of the host has a right-side cabinet door hinge. The middle of the host has a transmission gear set 13. The bottom of the main unit housing is provided with a bottom roller fixing shaft 8, a roller steering connection assembly 9, and a circuit module 14. The bottom roller fixing shaft 8 is used to fix the roller to the main unit housing.

[0033] like Figure 3 , 4 As shown, the software component is responsible for data processing and analysis. This component includes a scanning data acquisition module, a data preprocessing module 101, a three-dimensional comparison and analysis module 102, a data management and traceability module, and an equipment calibration and maintenance module. The scanning data acquisition module controls the working parameters and scanning process of the blue light scanner, enabling precise operation of the hardware. It interacts with the blue light scanner's hardware driver to control the hardware, sending control commands and receiving status information from the hardware, ensuring the smooth operation of the scanning process. The data analysis and detection software is responsible for in-depth analysis and processing of the received scanning data to achieve quality inspection of automotive parts. The data preprocessing module first accurately compares the scanned point cloud data with a pre-imported model, and then uses a specific algorithm to calculate the deviation between the actual size and the design size of the part, including shape deviation and dimensional deviation.

[0034] The scanning data acquisition module is primarily used to control the operating parameters and scanning process of the blue light scanner, enabling precise operation of the hardware. It allows setting the scanning frequency, selecting an appropriate frequency based on the complexity of the part and the required detection accuracy to balance scanning efficiency and data quality. Setting the exposure time is also crucial; adjusting the exposure time ensures the camera accurately captures the blue light signal reflected from the part's surface, avoiding overexposure or underexposure. When scanning parts with diverse colors and materials, such as automotive interior components, the exposure time needs to be finely adjusted according to their surface characteristics. The scanning control software also features a scanning path planning function, automatically generating the optimal scanning path based on the part's shape and size, ensuring that every area of ​​the part's surface is completely and accurately scanned. The software interacts with the blue light scanner's hardware driver to control the hardware, sending control commands and receiving status information from the hardware, ensuring a smooth scanning process. The data preprocessing module 101 is one of the core software components of the entire system. It is primarily responsible for in-depth analysis and processing of the received scanned data to achieve quality inspection of automotive parts. The software first precisely compares the scanned point cloud data with the pre-imported model. Through a specific algorithm, it calculates the deviation between the actual dimensions and design dimensions of the part, including shape deviations and dimensional deviations. For the inspection of automotive body panels, the software carefully compares the scanned data of the panel with the model to accurately identify potential problems such as localized deformation and dimensional deviations. Based on the comparison results, the software automatically generates detailed inspection reports, which include various inspection data of the part, deviation analysis charts, and conclusions regarding whether the part is qualified. These reports provide crucial information for quality control and production decisions in automotive manufacturing companies, helping them to promptly identify and resolve part quality issues, thereby improving product quality and production efficiency.

[0035] Image data acquired by a blue light scanner is first converted into point cloud data using a specific algorithm. During this process, deformable stripes in the image are decoded and analyzed to extract the three-dimensional coordinates of each pixel, thus generating point cloud data containing a large number of discrete points. Due to the potential influence of environmental noise and equipment errors during the actual scanning process, point cloud data often contains some noisy points and outliers, necessitating denoising. Common denoising methods include Gaussian filtering, which smooths the point cloud data and removes noise by weighted averaging of each point and its neighborhood. Bilateral filtering can also be used, which considers not only the spatial distance between points but also their attribute differences, effectively removing noise while preserving the detailed features of the point cloud. In addition, filtering operations such as voxel filtering are performed, which divides the point cloud space into small voxels and takes the average value of points within each voxel as a representative point, thereby reducing the amount of point cloud data and improving subsequent processing efficiency.

[0036] The three-dimensional comparison and analysis module 102 aligns and compares the preprocessed point cloud data with the original model of the automotive part. It employs the Iterative Closest Point (ICP) algorithm, continuously iterating to achieve the optimal spatial match between the point cloud data and the original model. During alignment, the distance from each point in the point cloud data to the surface of the original model is calculated, and this distance information is used to determine the deviation between the actual part and the design model. A specific algorithm calculates the deviation values ​​between the actual shape and the design shape of the part, such as dimensional deviations and shape deviations. For the inspection of automotive wheel hubs, the comparison analysis can accurately determine the deviations of parameters such as the diameter, spoke shape, and position of the wheel hub from the design values. Based on the set tolerance range, the module determines whether the part is qualified and generates a detailed inspection report. The report displays the deviation distribution in an intuitive chart format, providing accurate quality inspection results for automotive manufacturers.

[0037] like Figure 5 As shown, the non-contact automotive parts inspection system based on blue light scanning technology of the present invention includes the following process during operation: Scanning process: 1. Part Positioning and Fixing: Before scanning, the operator selects a suitable measuring bracket based on the shape, size, and structural characteristics of the automotive part. For parts with complex shapes, such as irregular sheet metal parts of a car body, a custom-made flexible fixture may be used, employing adjustable positioning blocks and a vacuum adsorption device to achieve stable fixation of the part. During the fixing process, the operator carefully checks the fit between the part and the positioning device to ensure that the part does not shift in any direction, thus guaranteeing the stability and accuracy of the part's position during subsequent scanning. 2. Scanner Startup and Scanning: After the parts are fixed in place, the operator starts the blue light scanner through the scanning control software. During startup, the software automatically initializes and checks the scanner's parameters to ensure they meet the requirements of the scanning task. The scanning frequency can be adjusted according to the complexity of the parts. For parts with simple shapes and few surface features, a higher scanning frequency can be selected to improve scanning efficiency; for parts with complex shapes and rich surface details, a lower scanning frequency is selected to ensure that subtle surface features are captured. The exposure time is also optimized according to the surface material and color of the parts. For highly reflective metal parts, the exposure time is shortened to avoid overexposure; for plastic parts with good light absorption, the exposure time is extended to ensure that the camera can accurately capture the blue light signal reflected from the part's surface. During scanning, the blue light scanner scans the parts from multiple angles and in all directions according to the preset scanning path. The scanning path is planned based on the part's 3D model to ensure that the entire surface of the part is covered and to avoid scanning blind spots. The scanner emits blue light from different angles and receives the reflected light, simultaneously capturing the blue light pattern reflected from the surface of the part. Using the principle of triangulation, it quickly calculates the three-dimensional coordinates of each point on the surface of the part, thereby obtaining complete surface data.

[0038] Data processing and detection workflow: 1. Data Transmission and Reception: After scanning, the large amount of point cloud data collected by the blue light scanner is quickly transmitted to the computer via a high-speed data transmission line and a report generation system. During transmission, the data undergoes verification and error correction to ensure its integrity and accuracy. The data analysis and detection software on the computer receives this data in real time and stores it in a designated memory area or hard disk space, awaiting further processing.

[0039] Algorithm 1: Noise Removal Algorithm: Eliminates interfering points and preserves true features. Statistical Outlier Removal (SOR) Algorithm: For each point cloud point, calculate the average distance of its "k neighboring points" (k is usually set to 20-50, adjusted according to the point cloud density), and then calculate the "standard deviation" based on the average distance of all points. Points with a distance greater than "average distance + 2-3 times the standard deviation" are identified as noise points and deleted.

[0040] Algorithm 2: Data Resampling Algorithm: Optimizes point cloud density and improves computational efficiency. Voxel Grid Downsampling Algorithm: The three-dimensional space is divided into uniformly sized "voxel grids" (the voxel size is set according to the detection accuracy, such as 0.05-0.1mm when the detection accuracy is ±0.1mm). For all point cloud points in each voxel, the average value of its three-dimensional coordinates is calculated. This average value is used to represent the points of the entire voxel, and other points in the voxel are deleted.

[0041] Algorithm 3: Point Cloud Segmentation Algorithm: Extracting Target Parts and Eliminating Background Interference First, select the "seed point" (usually a feature point on the surface of the target part, such as the center of a hole or the edge of the part), set the "growth conditions" (such as the angle between the normal vector of the adjacent point and the seed point being less than 10° and the distance being less than 2mm), and then start from the seed point to gradually include the neighboring points that meet the growth conditions into the target area until growth is no longer possible, and finally obtain the point cloud of the target part, while the background area is excluded.

[0042] 2. Analysis and Report Generation: The data analysis and inspection software first preprocesses the received data, including noise reduction and filtering, to remove interference data caused by environmental noise, equipment errors, and other factors, thus improving data quality. Next, the software accurately compares the preprocessed point cloud data with the pre-imported original model of the automotive part. Using a specific algorithm, it calculates the deviation between the actual dimensions and design dimensions of the part, including shape deviation and dimensional deviation. For automotive doors, the software analyzes in detail the deviations in the door's contour, edges, mounting holes, and other areas. Based on the comparison results, the software automatically identifies out-of-tolerance areas and marks them intuitively, such as using different colored blocks to represent different degrees of deviation. Finally, the software generates an inspection report containing the inspection results and deviation analysis. The report not only includes various inspection data and deviation analysis charts but also clearly states the conclusion of whether the part is qualified. These reports can be exported in common formats such as PDF and Excel, making it convenient for quality control personnel in automotive manufacturing companies to view and analyze, providing strong support for the company's production decisions.

[0043] Algorithm 1: Best Fit Alignment Algorithm, which extracts "key feature points" (such as hole center, part edge vertex, surface feature points) from the actual point cloud of the part, and at the same time extracts the corresponding "theoretical feature points" from the digital model (obtained through the original digital model feature extraction function). Construct an error function: With the goal of minimizing the sum of squared distances between actual feature points and theoretical feature points, calculate the optimal transformation matrix (including translations X / Y / Z and rotation angles α / β / γ) through iterative optimization (such as least squares method or singular value decomposition method).

[0044] Algorithm 2: Point-to-point deviation algorithm For each point in the actual point cloud, find the "nearest theoretical point" in the digital model (by calculating the spatial distance and selecting the digital model vertex or sampling point with the smallest distance), and calculate the Euclidean distance between the two points. This distance is the "point-to-point deviation" of the point. If the deviation is positive, it means that the actual point is "outside" the theoretical point in the digital model; if it is negative, it means that the actual point is "inside".

[0045] Algorithm 3: Deviation Statistical Analysis Algorithm, which performs statistical calculations on all local deviation data to generate key statistical indicators, including: 1 / Mean Deviation: The arithmetic mean of all deviation values, reflecting the "average offset direction" of the part as a whole from the digital model (e.g., a mean of +0.02mm means the part as a whole is slightly larger than the design size). 2 / Standard Deviation: Reflects the degree of dispersion of deviation values. The smaller the standard deviation, the better the consistency of part dimensions (e.g., the standard deviation of automobile engine pistons is usually required to be <0.008mm). 3 / Max Positive Deviation: The maximum value among all positive deviations, reflecting the "most prominent" area of ​​the part (such as a local bulge in a body panel). 4 / Max Negative Deviation: The minimum (largest absolute value) of all negative deviations, reflecting the "most recessed" area of ​​the part (such as shrinkage marks on injection molded parts). 5 / Out-of-Tolerance Rate: The proportion of out-of-tolerance points (absolute deviation value > upper tolerance limit) to the total number of points. The lower the out-of-tolerance rate, the higher the part pass rate (e.g., the out-of-tolerance rate requirement for key automotive parts is <0.5%).

[0046] This invention, a non-contact automotive parts inspection fixture based on blue light scanning technology, completely avoids physical contact with the part surface through non-contact blue light projection and dual-camera synchronous acquisition technology, fundamentally eliminating the risk of part damage and reducing the scrap rate of precision / vulnerable parts to below 0.1%. Simultaneously, the blue light can cover the entire part area in one pass (maximum scanning range up to 2000mm × 1500mm), and combined with a high-speed image sensor (acquisition frame rate ≥ 50fps), the inspection time for a single complex part is shortened to 5 minutes, improving efficiency by 5-15 times, significantly meeting the "mass production + rapid quality inspection" needs of the automotive manufacturing industry. This invention achieves breakthroughs through two major technological innovations: First, it employs a narrow-band blue light source of 450-470nm + an anti-reflective coating and adaptation algorithm, which not only reduces ambient light interference (measurement error is stably controlled within ±0.02mm), but also compensates for light absorption / reflection issues of dark / highly reflective parts through the algorithm, improving data integrity to over 99.8%; Second, it integrates an embedded high-speed data processing module, which can complete point cloud data stitching, digital-to-analog comparison, and deviation analysis in real time. From scanning to generating a visual inspection report (including deviation heatmap and out-of-tolerance point annotations), it only takes 3-5 minutes, achieving "results upon scanning," helping production lines adjust process parameters in real time and reducing the generation of batches of defective products. This invention has a three-in-one function of "dimensional measurement + appearance defect detection + assembly gap analysis": through high-precision point cloud data (point cloud density ≥1000 points / mm) 2 This system can simultaneously detect part dimensional tolerances (such as hole diameter and wall thickness), surface defects (such as scratches and dents), and assembly gaps (such as the fit gap between the door and the body) without changing equipment or adjusting tooling. Furthermore, the system supports seamless integration with automotive manufacturing MES (Production Execution System) and PLM (Product Lifecycle Management System), automatically synchronizing detection data to the enterprise management platform. This achieves a closed-loop process of "detection-analysis-process optimization," reducing equipment procurement costs by 40%-60% and improving data integration efficiency by over 80%.

[0047] The foregoing has shown and described the basic principles, main features, and advantages of the present invention. Those skilled in the art should understand that the present invention is not limited to the above embodiments. The embodiments and descriptions in the specification are merely illustrative of the principles of the invention. Various changes and modifications can be made to the invention without departing from its spirit and scope, and all such changes and modifications fall within the scope of the present invention as claimed. The scope of protection of this invention is defined by the appended claims and their equivalents.

Claims

1. A non-contact automotive parts inspection fixture system based on blue light scanning technology, characterized in that, The system comprises hardware and software components. The hardware component is responsible for data acquisition and transmission, while the software component handles data processing and analysis. The hardware component includes a blue light scanner and a device transmission and processing unit. The blue light scanner acquires three-dimensional information of a part's surface by emitting blue light and receiving reflected light. The blue light scanner includes a blue light emitting device and a high-resolution industrial camera. The blue light emitting device emits blue light of specific wavelengths and intensities to form a structured light pattern projected onto the part's surface. The high-resolution industrial camera simultaneously captures the blue light pattern reflected from the part's surface from different angles. The device transmission and processing unit includes a device transmission unit and a data processing unit. The device transmission unit is responsible for quickly and accurately transmitting the point cloud data acquired by the blue light scanner to the data processing unit. The data processing unit is a high-performance computer equipped with a powerful CPU, GPU, and large-capacity memory. The CPU is responsible for logical data processing and the execution of analysis algorithms. The GPU is mainly used to accelerate graphics processing and 3D model calculations, while large-capacity memory is used to store and quickly retrieve large amounts of scan data and intermediate calculation results. The software includes a scan data acquisition module, a data preprocessing module, a 3D comparison and analysis module, a data management and traceability module, and an equipment calibration and maintenance module. The scan data acquisition module controls the working parameters and scanning process of the blue light scanner, enabling precise operation of the hardware device. It interacts with the hardware driver of the blue light scanner to control the hardware device, send control commands, and receive status information from the hardware to ensure the smooth progress of the scanning process. The data analysis and detection software is responsible for performing in-depth analysis and processing of the received scan data to achieve quality inspection of automotive parts. The data preprocessing module first accurately compares the scanned point cloud data with the pre-imported model, and then calculates the deviation between the actual size and the design size of the part, including shape deviation and dimensional deviation, using a specific algorithm.

2. The non-contact automotive parts inspection system based on blue light scanning technology according to claim 1, characterized in that, The device's transmission unit uses high-speed data transmission interfaces, such as Ethernet and USB 3.0, to meet the real-time transmission requirements of large data volumes.

3. The non-contact automotive parts inspection system based on blue light scanning technology according to claim 1, characterized in that, The point cloud data collected by the blue light scanner is transmitted to the computer in real time via a data cable. The data processing software in the computer performs preprocessing operations such as noise reduction, filtering, and alignment on the data, and then compares and analyzes it with the model to finally generate a detection report.

4. The non-contact automotive parts inspection fixture system based on blue light scanning technology according to claim 1, characterized in that, The scanning control software sets the scanning frequency and selects an appropriate scanning frequency based on the complexity of the part and the required detection accuracy to balance scanning efficiency and data quality. By adjusting the exposure time, it ensures that the high-resolution industrial camera can accurately capture the blue light signal reflected from the surface of the part, avoiding overexposure or underexposure. When scanning automotive interior parts, which have diverse colors and materials, the exposure time needs to be finely adjusted according to the surface characteristics of the part. The scanning control software has a scanning path planning function, which automatically generates the optimal scanning path based on the shape and size of the part, ensuring that every area of ​​the part surface can be scanned completely and accurately.

5. The non-contact automotive parts inspection fixture system based on blue light scanning technology according to claim 1, characterized in that, The data analysis and testing software carefully compares the scanned data and models of automotive body panels to accurately identify potential local deformations and dimensional deviations. Based on the comparison results, the software automatically generates detailed inspection reports. These reports include various inspection data, deviation analysis charts, and qualification / disqualification conclusions for the parts. These reports provide important information for automotive manufacturers' quality control and production decisions, helping them to promptly identify and resolve part quality issues, thereby improving product quality and production efficiency.

6. The non-contact automotive parts inspection system based on blue light scanning technology according to claim 1, characterized in that, The blue light scanner is used for structured light projection and image acquisition. Specifically, the blue light scanner uses a specific optical device to project a carefully designed grating pattern onto the surface of an automotive part in the form of blue light. These grating patterns have periodic stripe structures, such as sinusoidal stripes and Gray code stripes. Taking sinusoidal stripes as an example, the brightness of the stripes changes according to a sine function. When blue light is projected onto the surface of the part, the originally regular grating pattern will be deformed due to the undulations of the part's surface. Two high-resolution industrial cameras simultaneously acquire these deformed stripe images from different angles. The position and angle of the high-resolution industrial cameras are precisely calibrated to ensure that the deformation information of each area of ​​the part's surface can be accurately captured. When scanning automotive engine blades, blue light is projected onto the blade surface. The complex curvature of the blade causes the grating pattern to produce unique deformations. Two high-resolution industrial cameras quickly capture these deformed stripes from different perspectives, providing raw image data for subsequent three-dimensional coordinate calculations.

7. The non-contact automotive parts inspection system based on blue light scanning technology according to claim 1, characterized in that, There is a fixed geometric relationship between the high-resolution industrial camera and the projector, including their baseline distance and relative angle. When blue light is projected onto the surface of the part and reflected, the pixels in the deformed fringe image received by the high-resolution industrial camera correspond to the original fringe projected by the projector. By analyzing these correspondences and using trigonometric functions and geometric relationships, the three-dimensional coordinates of each point on the surface of the part can be calculated. Assuming the optical center of the high-resolution industrial camera is O, the optical center of the projector is P, a point on the surface of the part is M, the baseline distance between the high-resolution industrial camera and the projector is b, and the imaging point of point M on the image plane captured by the high-resolution industrial camera is m, by analyzing the deformation of the fringe in the image, the angle information θ corresponding to point m can be obtained. According to the trigonometric relationship, in triangle OMP, given OP = b and θ, the distance z from point M to the plane where the high-resolution industrial camera and the projector are located can be calculated. Combining the imaging model of the high-resolution industrial camera and other geometric parameters, the x and y coordinates of point M in three-dimensional space can be further calculated, thereby achieving accurate measurement of the three-dimensional coordinates of each point on the surface of the part.

8. The non-contact automotive parts inspection fixture system based on blue light scanning technology according to claim 1, characterized in that, The image data acquired by the blue light scanner is first converted into point cloud data through a specific algorithm. During this process, the deformed stripes in the image are decoded and analyzed to extract the three-dimensional coordinate information corresponding to each pixel, thereby generating point cloud data containing a large number of discrete points. Due to the influence of environmental noise and equipment errors during the actual scanning process, there are often some noise points and outliers in the point cloud data. Therefore, it is necessary to perform noise reduction processing. Common noise reduction methods include Gaussian filtering, which smooths the point cloud data and removes noise by weighted averaging of each point and its neighborhood. Bilateral filtering can also be used, which not only considers the spatial distance of points but also the differences in point attributes, and can effectively remove noise while preserving the detailed features of the point cloud. In addition, filtering operations such as voxel filtering are also performed. This divides the point cloud space into small voxels and takes the average value of the points in each voxel as a representative point, thereby reducing the amount of point cloud data and improving the efficiency of subsequent processing.

9. The non-contact automotive parts inspection fixture system based on blue light scanning technology according to claim 1, characterized in that, The data analysis and inspection software aligns and compares the preprocessed point cloud data with the original model of the automotive parts. It employs the Iterative Closest Point (ICP) algorithm, continuously iterating to achieve the optimal spatial match between the point cloud data and the original model. During alignment, the distance from each point in the point cloud data to the surface of the original model is calculated. Based on this distance information, the deviation between the actual part and the design model is determined. A specific algorithm calculates the deviation values ​​between the actual shape and the design shape of the part, such as dimensional and shape deviations. For automotive wheel hub inspection, comparative analysis can accurately determine the deviations between the wheel hub's diameter, spoke shape, and position parameters and the design values. Based on the set tolerance range, the software determines whether the part is qualified and generates a detailed inspection report. The inspection report displays the deviation distribution in an intuitive chart format, providing accurate quality inspection results for automotive manufacturers.

10. The operation method of the non-contact automotive parts inspection tool system based on blue light scanning technology as described in claim 1, characterized in that, Specifically, the following steps are included: S1. Part Positioning and Fixing: Before scanning, the operator will select a suitable measuring bracket based on the shape, size and structural characteristics of the automotive parts. For parts with complex shapes, such as irregular sheet metal parts of the car body, a customized flexible fixture may be used. Adjustable positioning blocks and vacuum adsorption devices are used to achieve stable fixation of the parts. During the fixation process, the operator will carefully check the fit between the parts and the positioning device to ensure that the parts will not be displaced in any direction, so as to ensure the stability and accuracy of the part position during subsequent scanning. S2. Scanner Start-up and Scanning: After the parts are fixed in place, the operator starts the blue light scanner through the scanning control software. During the startup process, the software will automatically initialize and check the various parameters of the scanner to ensure that the parameter settings meet the requirements of this scanning task. The scanning frequency can be adjusted according to the complexity of the parts. For parts with simple shapes and few surface features, a higher scanning frequency can be selected to improve scanning efficiency; for parts with complex shapes and rich surface details, a lower scanning frequency should be selected to ensure that the subtle features of the part surface can be captured. The exposure time will also be optimized according to the surface material and color of the parts. For metal parts with strong reflectivity, the exposure time should be shortened appropriately to avoid overexposure. For plastic parts with good light absorption, the exposure time is appropriately extended to ensure that the camera can accurately capture the blue light signal reflected from the surface of the part. During the scanning process, the blue light scanner scans the part from multiple angles and in all directions according to the preset scanning path. The scanning path is planned based on the three-dimensional model of the part to ensure that the entire surface of the part is covered and to avoid scanning blind spots. The scanner emits blue light from different angles and receives reflected light, and simultaneously captures the blue light pattern reflected from the surface of the part. Using the principle of triangulation, the three-dimensional coordinates of each point on the surface of the part are quickly calculated to obtain complete surface data. S3. Data transmission and reception: After scanning, the large amount of point cloud data collected by the blue light scanner will be transmitted to the computer quickly through a high-speed data transmission line and a report generation system. During the transmission process, the data will be verified and error corrected to ensure the integrity and accuracy of the data. The data analysis and detection software in the computer will receive the data in real time and store it in a designated memory area or hard disk space. S4. Analysis and Report Generation: The data analysis and inspection software first preprocesses the received data, including noise reduction and filtering, to remove interference data caused by environmental noise and equipment errors, thus improving data quality. Next, the software accurately compares the preprocessed point cloud data with the pre-imported original model of the automotive part. Using a specific algorithm, it calculates the deviation between the actual size and design size of the part, including shape and dimensional deviations. For automotive doors, the software analyzes in detail the deviations in the door's contour, edges, mounting holes, etc. Based on the comparison results, the software automatically identifies out-of-tolerance areas and marks them intuitively, such as using different colored blocks to represent different degrees of deviation. Finally, the software generates an inspection report containing the inspection results and deviation analysis. The report not only includes various inspection data and deviation analysis charts but also clearly states whether the part is qualified. These reports can be exported in common formats such as PDF and Excel, making it convenient for quality control personnel in automotive manufacturing companies to view and analyze, providing strong support for the company's production decisions.

11. The operation method of the non-contact automotive parts inspection tool system based on blue light scanning technology according to claim 10, characterized in that, The specific algorithm used in step S3 includes: Algorithm 1: Noise Removal Algorithm: Eliminates interfering points and preserves true features. The statistical filtering algorithm Statistical Outlier Removal (SOR) calculates the average distance k of the k nearest neighbors for each point in the point cloud. The k value is usually set to 20-50 and adjusted according to the point cloud density. Then, it calculates the standard deviation based on the average distance of all points and identifies and deletes points whose distance is greater than the average distance + 2-3 times the standard deviation as noise points. Algorithm 2: Data Resampling Algorithm: Optimizes point cloud density and improves computational efficiency. The VoxelGrid Downsampling algorithm divides the three-dimensional space into uniformly sized "voxel grids". The voxel size is set according to the detection accuracy. For example, when the detection accuracy is ±0.1mm, the voxel size is set to 0.05-0.1mm. For all point cloud points in each voxel, the average value of its three-dimensional coordinates is calculated. This average value is used to represent the points of the entire voxel, and other points in the voxel are deleted. Algorithm 3: Point Cloud Segmentation Algorithm: Extract the target part and eliminate background interference. First, select seed points, which are usually feature points on the surface of the target part, such as the center of a hole or the edge of the part. Set growth conditions, such as the angle between the normal vector of the adjacent point and the seed point is less than 10° and the distance is less than 2mm. Then, starting from the seed point, gradually include the neighboring points that meet the growth conditions into the target area until no more growth is possible. Finally, the point cloud of the target part is obtained, and the background area is excluded.

12. The operation method of the non-contact automotive parts inspection tool system based on blue light scanning technology according to claim 10, characterized in that, The specific algorithm mentioned in step S4 includes: Algorithm 1: Best Fit Alignment Algorithm. This algorithm extracts key feature points from the actual point cloud of the part, such as hole centers, part edge vertices, and surface feature points. Simultaneously, it extracts corresponding theoretical feature points from the digital model. The algorithm obtains these features through the original digital model feature extraction function and constructs an error function. The objective is to minimize the sum of squared distances between actual and theoretical feature points. The algorithm iteratively optimizes the transformation matrix using methods such as least squares and singular value decomposition. The optimal transformation matrix includes translations X / Y / Z and rotation angles α / β / γ. Algorithm 2: Point-to-point deviation algorithm. For each point in the actual point cloud, find the nearest theoretical point in the digital model. Through spatial distance calculation, select the digital model vertex or sampling point with the smallest distance and calculate the Euclidean distance between the two points. This distance is the point-to-point deviation of the point. If the deviation is positive, it means that the actual point is outside the theoretical point in the digital model; if it is negative, it means that the actual point is inside. Algorithm 3: Deviation Statistical Analysis Algorithm, which performs statistical calculations on all local deviation data to generate key statistical indicators, including: 1 / Mean Deviation: The arithmetic mean of all deviation values, reflecting the average offset direction of the part from the digital model. For example, a mean of +0.02mm indicates that the part is slightly larger than the design size. 2 / Standard Deviation: Reflects the degree of dispersion of deviation values. The smaller the standard deviation, the better the consistency of part dimensions. For example, the standard deviation of the piston of an automobile engine is usually required to be <0.008mm. 3 / Max Positive Deviation: The maximum value among all positive deviations, reflecting the most prominent area of ​​the part, such as a local bulge in a body panel. 4 / Max Negative Deviation: The absolute value of the minimum of all negative deviations is the largest, reflecting the most concave area of ​​the part, such as shrinkage marks in injection molded parts; 5 / Out-of-Tolerance Rate: The proportion of out-of-tolerance points to the total number of points. The absolute value of the deviation of out-of-tolerance points is greater than the upper limit of tolerance. The lower the out-of-tolerance rate, the higher the pass rate of the parts. For example, the out-of-tolerance rate requirement for key automotive parts is <0.5%.

Images