Visual detection method and system for inner hole threads of prestressed anchorage device clamping piece

Abstract

The invention relates to the technical field of anchorage device production detection, in particular to a visual detection method and system for an inner hole thread of a prestressed anchorage device clamping piece, and the method comprises the steps: carrying out the structural region division of the inner hole thread of the prestressed anchorage device clamping piece to be detected in the axial direction, and obtaining at least three thread detection regions; performing in-hole image acquisition on each thread detection area to obtain a corresponding area image, and obtaining a corresponding acquisition condition parameter set; performing geometric feature extraction on the region image to obtain a geometric feature parameter set corresponding to each thread detection region; for each thread detection region, inputting the geometric feature parameter set into the thread defect feature evaluation model corresponding to the region to obtain a region defect score; correcting the regional defect score by considering the error influence of the acquisition condition parameter set on the actual characteristics; and carrying out coupling analysis on all the corrected regional defect scores to obtain a comprehensive thread defect score of the pre-stressed anchorage device clamping piece to be detected.

Description

Technical Field

[0001] The present invention relates to the technical field of anchor production detection, and in particular to a method and system for visually detecting threads in inner holes of prestressed anchor clips. Background Art

[0002] In major infrastructure projects such as bridge engineering, geotechnical anchoring, and high-rise buildings, prestressed anchors serve as core components that bear the weight of the structure. The machining quality of the internal threads of the clips directly determines the mechanical engagement performance and long-term service reliability of the anchor system. However, due to the miniaturization of thread structures and the complexity of machining processes, hidden defects such as shallow bottom threads, cumulative pitch deviation, and asymmetric thread angles are prone to occur during the production process. With the development of automated inspection technology, thread inspection methods based on machine vision are gradually gaining application.

[0003] However, existing visual inspection methods for internal threads mainly determine whether the threads are qualified by acquiring images of the internal threads and then comparing them with standard values based on preset single feature parameters (such as thread major diameter, pitch, etc.). However, existing single-feature inspection methods have obvious limitations: First, thread quality is a multi-dimensional comprehensive indicator, and it is difficult to fully and accurately evaluate the actual quality of the threads by relying solely on a single feature parameter. For example, threads with qualified pitch but abnormal tooth angles will also cause problems during assembly and use; second, single-feature inspection methods lack targeted evaluation of different inspection areas. Different parts of the internal thread of the prestressed anchor clip differ in function and stress conditions, and have different quality requirements. Existing methods often evaluate the entire thread as a whole, failing to highlight key quality issues in each part. Summary of the Invention

[0004] In order to solve the above technical problems, the present invention provides a method and system for visually detecting the inner hole threads of prestressed anchor clips, which can improve the comprehensiveness, accuracy and pertinence of the detection of the inner hole threads of prestressed anchor clips.

[0005] In a first aspect, the present invention provides a method for visually inspecting threads in a prestressed anchor clip, the method comprising: Divide the inner hole thread of the prestressed anchor clip to be tested into structural areas along the axial direction to obtain at least three thread detection areas; Performing in-hole image acquisition on each of the thread detection areas to obtain a corresponding area image and a corresponding acquisition condition parameter set; Performing geometric feature extraction on the area image to obtain a geometric feature parameter set corresponding to each thread detection area; For each thread inspection area, inputting the geometric feature parameter set into the thread defect feature evaluation model corresponding to the area to obtain a regional defect score; Taking into account the error influence of the acquisition condition parameter set on the actual features, the regional defect score is corrected; A coupling analysis is performed on all the corrected regional defect scores to obtain a comprehensive thread defect score of the prestressed anchor clip to be tested, which is then compared with a preset thread defect threshold to generate a test result.

[0006] Further, in response to the comprehensive thread defect score being higher than a preset thread defect threshold; For each of the thread detection areas, a deviation calculation is performed between the corresponding geometric feature parameter set and the design requirements of the inner hole thread of the prestressed anchor clip to be tested, so as to obtain a geometric feature deviation vector corresponding to the thread detection area; According to the positional relationship of the thread axis, all the geometric feature deviation vectors are serially adjusted to obtain the internal hole thread feature deviation matrix; The internal hole thread feature deviation matrix is input into a preset defect identification and tracing model, and the defect type and the process parameter adjustment strategy corresponding to the defect are output.

[0007] Furthermore, the internal hole thread feature deviation matrix is: Among them, d nm Indicates the deviation value between the mth geometric feature parameter of the nth thread detection area and the design requirements.

[0008] Furthermore, the thread detection area includes a thread starting section, a thread middle section and a thread ending section.

[0009] Furthermore, the acquisition condition parameter set includes light intensity, shooting angle and device vibration intensity.

[0010] Furthermore, the geometric feature parameter set includes thread major diameter, thread minor diameter, thread pitch, thread profile angle and thread profile curvature.

[0011] Furthermore, considering the error influence of the acquisition condition parameter set on the actual features, the regional defect score is corrected, and the correction steps include illumination intensity correction, shooting angle correction, equipment vibration intensity correction, geometric feature extraction error correction and environmental factor correction.

[0012] On the other hand, the present application also provides a prestressed anchor clip inner hole thread visual inspection system, the system comprising: A region division module divides the internal thread of the prestressed anchor clip to be tested into structural regions along the axial direction to obtain at least three thread detection regions, wherein the thread detection regions include a thread starting section, a thread middle section, and a thread ending section; An image acquisition module is configured to acquire an in-hole image of each thread detection area, obtain a corresponding area image, and obtain a corresponding acquisition condition parameter set, wherein the acquisition condition parameter set includes light intensity, shooting angle, and equipment vibration intensity; a feature extraction module, performing geometric feature extraction on the regional image to obtain a geometric feature parameter set corresponding to each thread detection region, wherein the geometric feature parameter set includes thread major diameter, thread minor diameter, thread pitch, thread profile angle, and thread profile curvature; a defect feature evaluation module, for each of the thread inspection areas, inputting the geometric feature parameter set into a thread defect feature evaluation model corresponding to the area to obtain an area defect score; a scoring correction module, which corrects the regional defect score by taking into account the error influence of the acquisition condition parameter set on the actual feature; The comprehensive analysis module performs a coupling analysis on all the corrected regional defect scores to obtain a comprehensive thread defect score of the prestressed anchor clip to be tested, and compares it with a preset thread defect threshold to generate a test result.

[0013] In a third aspect, the present application provides an electronic device comprising a bus, a transceiver, a memory, a processor, and a computer program stored in the memory and runnable on the processor, wherein the transceiver, the memory, and the processor are connected via the bus, and the computer program implements the steps of any one of the above methods when executed by the processor.

[0014] In a fourth aspect, the present application further provides a computer-readable storage medium having a computer program stored thereon, which implements the steps in any one of the above methods when executed by a processor.

[0015] Compared with existing technologies, the present invention has the following advantages: This method simultaneously extracts multi-dimensional geometric characteristic parameters, including the major diameter, minor diameter, pitch, thread angle, and thread profile curvature, covering the core quality indicators of thread shape, size, and profile. It can identify hidden defects such as "qualified pitch but abnormal thread angle," avoiding misjudgment based on a single parameter. Multi-dimensional characteristics are directly related to the mechanical engagement performance of the thread. By integrating multiple characteristic parameters into the defect evaluation model, it can more accurately reflect the reliability of the thread in actual service, rather than relying solely on a single dimensional standard. The internal thread is divided into at least three regions along the axial direction: the starting section, the middle section, and the ending section. Different regions have different functions in anchor assembly and have different quality requirements. Traditional methods evaluate the thread as a whole and cannot distinguish the impact of defects in different parts. This method uses independent scoring of different regions to accurately locate the specific area where defects occur, making it easier for production to trace processing problems and improve quality improvement efficiency. During image acquisition, environmental factors such as light intensity, shooting angle, and equipment vibration can cause errors in characteristic parameter measurement. This method records real-time environmental data through an acquisition condition parameter set and corrects regional defect scores based on an error model, reducing the impact of accidental factors on the inspection results. Traditional methods ignore the influence of acquisition conditions, which may cause fluctuations in inspection results for the same thread under different environments. This method uses an error correction mechanism to ensure that the inspection results only reflect the actual processing quality of the thread, rather than accidental fluctuations in the equipment or environment, thereby improving the repeatability and reliability of the inspection. By integrating the corrected defect scores of each region through coupled analysis, not only the quality status of a single region is considered, but also the synergistic impact of defects between regions can be identified. The resulting comprehensive thread defect score can be directly compared with the preset threshold, and the comparison results can be quickly output, facilitating real-time sorting and quality control on the production line, thereby improving inspection efficiency. Each step of the method can be integrated into automated testing equipment to adapt to the production line rhythm, reduce manual intervention, and lower testing costs. The thread defect feature evaluation model can be continuously trained and optimized based on historical testing data to adapt to the testing needs of anchor clips of different specifications, improve the system's ability to identify complex defects, and has the potential for long-term technological iteration. This method breaks through the limitations of traditional single feature detection through multi-dimensional feature fusion, regional differentiated detection, environmental error correction and systematic comprehensive evaluation, and significantly improves the comprehensiveness, accuracy and pertinence of thread detection of inner holes in prestressed anchor clips. BRIEF DESCRIPTION OF THE DRAWINGS

[0016] Figure 1 Flowchart of a method for visually inspecting threads in inner holes of prestressed anchor clips according to an embodiment of the present invention; Figure 2 It is a structural diagram of the visual inspection system for the inner hole thread of the prestressed anchor clip in an embodiment of the present invention. DETAILED DESCRIPTION

[0017] The present application is described below in conjunction with the accompanying drawings.

[0018] like Figure 1 As shown, a method for visually inspecting the inner thread of a prestressed anchor clip of the present invention specifically comprises the following steps: S1. Divide the inner hole thread of the prestressed anchor clip to be tested into structural areas along the axial direction to obtain at least three thread detection areas, wherein the thread detection areas include a thread starting section, a thread middle section, and a thread ending section; S2. Capturing an in-hole image of each thread detection area to obtain a corresponding area image and a corresponding acquisition condition parameter set, wherein the acquisition condition parameter set includes light intensity, shooting angle, and equipment vibration intensity; S3, extracting geometric features from the regional image to obtain a set of geometric feature parameters corresponding to each thread detection area, wherein the set of geometric feature parameters includes thread major diameter, thread minor diameter, thread pitch, thread profile angle, and thread profile curvature; S4. For each thread inspection area, input the geometric feature parameter set into a thread defect feature evaluation model corresponding to the area to obtain a regional defect score; S5. Considering the error influence of the acquisition condition parameter set on the actual features, correct the regional defect score; S6. Perform coupling analysis on all the corrected regional defect scores to obtain a comprehensive thread defect score of the prestressed anchor clip to be tested, and compare it with a preset thread defect threshold to generate a test result.

[0019] In this embodiment, the method constructs a systematic quality assessment system by dividing thread detection into multiple steps and sequentially performing area division, image acquisition, feature extraction, defect evaluation, score correction and comprehensive analysis; it can ensure a comprehensive assessment of thread quality from multiple dimensions and multiple levels, avoiding the one-sidedness of single feature evaluation, thereby improving the accuracy and reliability of the assessment; by dividing the internal hole thread into multiple detection areas along the axial direction and independently evaluating each area, the method can accurately locate the areas where quality problems may exist in the thread; it enables producers to improve and optimize specific areas, thereby improving production efficiency and product quality; the method records the acquisition condition parameter set while acquiring the regional image, and considers the error influence of these parameters on the actual features in subsequent steps, and makes an accurate assessment of regional defects. The defect scores are corrected; it can significantly improve the accuracy of defect scores, reduce misjudgments caused by changes in acquisition conditions, and thus improve the reliability of the entire detection system; by coupling analysis of all corrected regional defect scores, the method can obtain the comprehensive thread defect score of the prestressed anchor clip to be tested, and compare it with the preset thread defect threshold to generate the detection result; thereby providing producers with comprehensive quality information to help them make more informed decisions; the overall logical design of the method has high adaptability and scalability; with the continuous advancement of detection technology and the improvement of anchor design requirements, it is easy to adjust the area division method, add new feature parameters, optimize the defect feature evaluation model or adjust the comprehensive analysis strategy to adapt to new detection needs, and can maintain high efficiency and accuracy in different scenarios.

[0020] In some embodiments of the present invention, the inner hole thread of the prestressed anchor clip to be tested is divided into structural areas along the axial direction to obtain at least three thread detection areas, wherein the thread detection areas include a thread starting section, a thread middle section, and a thread ending section; The internal thread of the prestressed anchor clip to be tested is divided into structural areas along the axial direction. Different parts of the internal thread of the prestressed anchor clip have different functions and stress conditions. The starting and ending sections of the thread are subject to greater stress concentration, while the middle section mainly bears stable axial tension. The difficulty of machining different areas may also vary, resulting in different probabilities and types of defects. The starting and ending sections are more likely to have shallow bottom teeth or abnormal tooth angles due to the entry and exit of the machining tool. Determine the starting, middle, and ending points of the internal thread according to the design drawings and processing requirements of the prestressed anchor clip. These points should be based on the actual function and stress conditions of the thread, as well as key nodes in the processing process. Divide the internal thread into at least three inspection areas, including the thread start segment, the thread middle segment, and the thread end segment. When dividing, ensure that the length of each segment is reasonable, which can fully reflect the characteristics of the area and facilitate subsequent image acquisition and feature extraction. Uniquely identify each divided thread inspection area and record the location of the division point and the length of each segment. Use high-precision measuring tools to measure the internal thread and determine the specific positions of the starting point, middle point and end point; According to the measurement results, make corresponding marks on the internal hole thread and use fixtures or tools to assist in area division; Record the division results and related information and keep them on file for future reference; The starting section of the thread first contacts the steel strand during tensioning, and is responsible for initial engagement and stress transfer, requiring high wear resistance and precision. It may also have defects such as tooth angle deviation, excessive surface roughness, and shallow bottom teeth. The middle section of the thread is used to withstand continuous axial tension and is the core area where stress is evenly distributed. It has defects such as cumulative pitch deviation and abnormal curvature of the tooth profile. The thread termination section is at the end of the anchoring, which is prone to defects due to process fluctuations when the tool withdraws, affecting the overall anchoring stability; it has defects such as sudden pitch change, tooth shape deformation, and insufficient local bottom diameter.

[0021] In this embodiment, the threaded portion of the prestressed anchor clip is divided into a starting section, an intermediate section, and an ending section based on the functions and stress differences of different sections. This overcomes the limitations of traditional overall testing. This division method allows the testing process to closely match the actual working conditions of the thread. For example, a specialized test is performed on the starting section, which is subject to initial engagement and stress transfer and requires high wear resistance and precision. This effectively avoids the problem of missed detection of defects in key areas due to general testing methods, and significantly improves the accuracy and reliability of testing. Taking into account the differences in machining difficulty, defect probability, and type across different regions, this division method allows for focused inspection of typical defects in each region. For example, differentiated inspection strategies are developed for problems such as shallow bottom teeth and abnormal tooth angles that are prone to occur in the starting and ending sections due to tool entry and exit, as well as for cumulative pitch deviations in the middle section. This improves the efficiency of identifying hidden and regional defects. Compared to traditional inspection methods, this method can locate defects more quickly and accurately, saving significant time and cost for subsequent repairs and process improvements. The division points are determined strictly according to the design drawings and processing requirements of the prestressed anchor clips, combined with the actual function of the thread, stress conditions and key processing nodes. The specific positions are determined using high-precision measurement tools to ensure scientific and accurate regional division. The reasonable regional length setting fully reflects the characteristics of each area and facilitates subsequent image acquisition and feature extraction, making each link of the inspection process closely connected and efficient, and improving the standardization and professionalism of the entire inspection process. Each thread inspection area after division is uniquely identified, and the location of the division point, the length of each section and other information are recorded in detail and archived, providing complete and accurate data support for quality traceability; once a quality problem is found, the area where the defect is located can be quickly identified, and the cause can be analyzed based on the regional characteristics. At the same time, it provides a strong reference for the improvement of the processing technology, forming a virtuous cycle of quality improvement and promoting the continuous improvement of the production and manufacturing level of prestressed anchor clips.

[0022] In some embodiments of the present invention, an in-hole image is captured for each of the thread detection areas to obtain a corresponding area image and a corresponding acquisition condition parameter set; Industrial-grade line scan cameras paired with telecentric lenses solve the perspective distortion problem caused by deep holes and small diameters of internal threads, ensuring undistorted tooth profile imaging. For tiny defect detection, high-resolution area scan cameras can be used to achieve sub-pixel edge detection. Adopt a multi-light source combination scheme, including an annular coaxial light source, a side strip light source and a controllable optical fiber bundle; Equipped with a high-precision linear module and a rotating platform, the camera achieves uniform scanning along the thread axis and 360° circumferential shooting, ensuring coverage of the entire thread inspection area. For long thread clips, segmented stitching imaging technology can be used to avoid insufficient field of view in a single imaging shot. Fix the prestressed anchor clip to be tested on the test bench and use a positioning device to ensure the alignment accuracy between the inner hole thread and the camera; According to the characteristics of the thread inspection area and the inspection requirements, set the camera's exposure time, gain and other parameters, as well as the light source's brightness and color temperature. Each thread inspection area is photographed from multiple angles to obtain thread images from different perspectives, which can reflect the geometric characteristics and potential defects of the thread; In order to improve the reliability and accuracy of image data, multiple images can be collected for each inspection area, and the image with the best quality can be selected as the basis for subsequent processing; The acquisition condition parameter set includes light intensity, shooting angle and device vibration intensity; Record the light intensity value during each image acquisition to facilitate subsequent analysis of the impact of lighting conditions on image quality and make corrections when necessary; Record the angle information of each shot, including pitch angle, yaw angle, etc., so as to accurately restore the three-dimensional shape of the thread in subsequent processing; Use devices such as vibration sensors to record the vibration intensity of the equipment during image acquisition, evaluate the impact of vibration on image quality, and take vibration reduction measures when necessary.

[0023] In this embodiment, an industrial-grade line array camera paired with a telecentric lens effectively solves the perspective distortion problem caused by deep holes and small diameters of internal threads, ensuring undistorted tooth profile imaging and providing a high-quality image foundation for subsequent geometric feature extraction and defect evaluation. For tiny defect detection, a high-resolution area array camera is used to achieve sub-pixel edge detection, further improving the precision and accuracy of imaging. The multi-light source combination solution, including an annular coaxial light source, a side strip light source, and a controllable fiber optic bundle, provides flexible and variable lighting conditions, helping to fully capture the geometric characteristics and potential defects of the thread. Equipped with a high-precision linear module and a rotating platform, the camera achieves uniform scanning along the thread axis and 360° circumferential shooting, ensuring coverage of the entire thread inspection area and avoiding blind spots. For long thread clips, the use of segmented splicing imaging technology effectively solves the problem of insufficient field of view in a single imaging, improving the comprehensiveness and reliability of inspection. Based on the characteristics and inspection requirements of the thread inspection area, the camera's exposure time, gain, and other parameters, as well as the brightness and color temperature of the light source, are carefully set to ensure optimal conditions for image acquisition. Each inspection area is photographed from multiple angles to obtain thread images from different perspectives, providing rich data support for subsequent 3D morphological restoration and defect analysis. The acquisition condition parameter set, including light intensity, shooting angle, and equipment vibration intensity, is recorded for each image acquisition. This provides a basis for subsequent analysis of the impact of lighting conditions on image quality, accurate restoration of the 3D morphology of the thread, and assessment of the impact of vibration on image quality. Correction and vibration reduction measures are then implemented when necessary. Through automated image acquisition processes and parameter settings, manual intervention is reduced and detection efficiency is improved; recording detailed acquisition condition parameter sets makes it possible to optimize and automate subsequent detection processes, promoting the automation process in the field of anchor production inspection.

[0024] More specifically, geometric feature extraction is performed on the regional image to obtain a geometric feature parameter set corresponding to each thread detection area, wherein the geometric feature parameter set includes thread major diameter, thread minor diameter, thread pitch, thread profile angle, and thread profile curvature; To address possible noise in the captured images, such as equipment vibration and uneven lighting, median filtering or Gaussian filtering is used to eliminate random noise. For noise that is prone to appearing on thread edges, bilateral filtering is used to smooth out the noise while preserving edge details. Use histogram equalization or adaptive histogram equalization to improve the contrast between the thread profile and the background. To address the shadow problem in the deep hole area of the internal thread, use homomorphic filtering to separate the illumination and reflection components to enhance local details. Based on the camera calibration parameters, the distortion correction algorithm is used to perform geometric transformation on the image to eliminate the deformation of the thread profile caused by lens distortion and ensure the accuracy of subsequent measurements; The thread contour is determined based on the edge detection results. The farthest and closest points of the contour are found along the thread axis within each detection area, which are used as the estimated values of the major and minor diameters of the thread, respectively. The contour is fitted to obtain more accurate major and minor diameter values. In the axial direction of the thread, select a certain number of adjacent tooth tops or tooth bottoms; calculate the distance between these points and take the average value as the estimated value of the pitch; On the thread profile, select two adjacent flanks; calculate the angle between these two edges as the estimated value of the thread angle; to improve accuracy, the thread angle can be measured at multiple locations and the average value can be taken; Perform curvature analysis on the thread profile; use a curvature calculation algorithm to obtain the curvature value of each point on the thread profile; analyze the curvature change trend to evaluate the symmetry and smoothness of the thread profile; The extracted thread major diameter, thread minor diameter, pitch, tooth profile angle and tooth profile curvature parameters are combined into a geometric feature parameter set.

[0025] In this embodiment, by adopting algorithms such as median filtering, Gaussian filtering, and bilateral filtering, random noise in the image and noise at the edge of the thread are effectively eliminated, thereby improving the clarity and processing accuracy of the image. By using technologies such as histogram equalization, adaptive histogram equalization, and homomorphic filtering, the contrast between the thread profile and the background is enhanced, the shadow problem in the deep hole area of the internal thread is solved, and the thread features are made more obvious, which facilitates subsequent feature extraction. Based on camera calibration parameters, a distortion correction algorithm is used to perform geometric transformation on the image, eliminating thread profile deformation caused by lens distortion, ensuring the accuracy of subsequent measurements and providing a foundation for obtaining precise geometric feature parameters. Through edge detection, contour fitting and other technologies, key geometric feature parameters such as the major diameter, minor diameter, and pitch of the thread are accurately extracted, providing an important basis for evaluating thread quality. The extraction of the tooth profile angle and tooth profile curvature further enriches the dimensions of thread quality assessment, making the evaluation results more comprehensive and accurate. The extracted parameters such as thread major diameter, thread minor diameter, pitch, tooth profile angle and tooth profile curvature are combined into a geometric feature parameter set, providing reliable input data for the subsequent thread defect feature evaluation model; The geometric feature parameters extracted in this step are universal and can adapt to the detection needs of the internal thread of prestressed anchor clips of different types and specifications; by adjusting the image processing algorithm and parameter settings, the detection effect can be further optimized and the adaptability and flexibility of the detection method can be improved.

[0026] Furthermore, for each of the thread inspection areas, the geometric feature parameter set is input into a thread defect feature evaluation model corresponding to the area to obtain an area defect score; The thread defect feature evaluation model is a regional customized model. The model of the thread starting section adds an imported taper deviation detection item; the thread ending section model includes a thread end transition smoothness evaluation item; the thread middle section model sets the pitch cumulative error tolerance threshold; The thread start segment is the first part of the anchor clip that contacts the prestressed tendon. The accuracy of its lead-in taper is crucial to ensuring mechanical engagement performance. Therefore, a test item for lead-in taper deviation is added to the thread defect characteristic evaluation model for the start segment. By comparing the difference between the actual lead-in taper and the standard taper, the machining accuracy of the start segment thread can be evaluated. The thread termination model includes an evaluation item for thread end transition smoothness. This evaluation item is crucial for preventing stress concentration and ensuring long-term service reliability. By analyzing the profile changes at the thread end, the smoothness of the transition can be assessed, thereby determining whether there are potential defects. A tolerance threshold for cumulative pitch error is set in the thread mid-segment model. As the primary load-bearing component of the anchor clip, the accuracy of its pitch is crucial for ensuring mechanical engagement and long-term service reliability. However, due to processing limitations, the pitch may have a certain degree of cumulative error. Therefore, a tolerance threshold for cumulative pitch error is set in the mid-segment model. A defect is only identified when the cumulative pitch error exceeds this threshold. Input the extracted geometric feature parameter set of each thread inspection area into the corresponding thread defect feature evaluation model; Each model calculates the defect score of the thread inspection area based on the input geometric feature parameter set and area-customized evaluation items.

[0027] In this embodiment, by setting specific evaluation items and tolerance thresholds for different thread inspection areas, the model can more comprehensively consider multiple dimensions of thread quality, thereby more accurately assessing the actual quality status of each area. This avoids the limitations of traditional single feature parameter evaluation methods and can capture more potential defects and processing errors. The thread start segment model adds a lead-in taper deviation detection item to ensure that the first part of the anchor clip in contact with the prestressed tendon has an accurate lead-in taper, thereby ensuring mechanical engagement performance. The thread end segment model includes a thread end transition smoothness evaluation item to help prevent stress concentration and improve the long-term service reliability of the anchor system. The thread middle segment model sets a pitch cumulative error tolerance threshold, taking into account the limitations of the processing technology while ensuring the pitch accuracy of the main load-bearing part. The regional customized model makes quality control more refined and can carry out targeted improvement and optimization for key quality issues in different regions. By providing an independent defect score for each region, it is easier to identify areas with weak quality and take corresponding measures to improve them. The regional customized thread defect feature evaluation model can be integrated into the automated inspection system to achieve fast and accurate assessment of thread quality.

[0028] In some embodiments of the present invention, the regional defect score is corrected by taking into account the error influence of the acquisition condition parameter set on the actual feature; Light intensity correction: Changes in light intensity may cause changes in image contrast, detail, and edge clarity, thereby affecting the extraction of geometric features. Based on the different lighting conditions of multiple image acquisitions, an illumination balancing algorithm is used to pre-process the image to eliminate the impact of light intensity fluctuations. By comparing the detection results under different lighting conditions, the degree of influence of light intensity on the score can be determined, and a correction coefficient can be used to adjust the score. Shooting angle correction: Changes in shooting angle may cause distortion in the geometric features of the thread. Especially for larger threads or high-precision requirements, slight angle changes may also introduce errors. By analyzing images taken at different shooting angles and using computer vision technology, the images are geometrically corrected. The relationship between shooting angle and feature error can be established by using model images taken at known standard angles. By compensating for shooting angle errors, the score can be corrected. Correcting equipment vibration intensity: Equipment vibration can cause image blur, blurring the edges of threads and thus affecting geometric feature extraction. Equipment vibration is monitored and modeled to assess the impact of vibration intensity on image clarity and geometric feature extraction. When vibration is high, image denoising algorithms can be used to reduce vibration-induced noise, and image stabilization techniques can be used to restore blurred images. Geometric feature extraction error correction: During image processing, geometric feature extraction may be affected by image quality, leading to feature extraction errors. By setting a standardized geometric feature extraction algorithm and using reference samples for error compensation, a feature extraction error model is established under different acquisition conditions. The extracted geometric features are corrected according to the error impact of each acquisition condition. Environmental factor correction: Environmental factors such as temperature and humidity may cause changes in the device's operating state, affecting the stability of image acquisition. By installing temperature and humidity sensors, changes in environmental parameters are monitored. Based on the impact of temperature and humidity changes on device performance, an environmental compensation model is established to adjust for errors caused by environmental changes. A single correction method may not be able to fully resolve all error problems; a multi-level correction strategy is adopted to integrate the correction results of lighting, shooting angle, equipment vibration, etc., comprehensively consider the impact of various factors on the final score, and generate a comprehensive correction coefficient; a final correction is performed on the regional defect score to ensure the accuracy and reliability of the score.

[0029] In this embodiment, by correcting parameters such as light intensity, shooting angle, and equipment vibration, the impact of these environmental and operating conditions on image quality is reduced, ensuring more accurate extraction of thread geometric features; enhancing detection accuracy: using illumination balancing algorithms, geometric correction, denoising technology, and other means to improve image clarity and feature extraction accuracy, especially under complex acquisition conditions, to ensure the accuracy of detection results; by combining the correction of environmental factors and multi-level correction strategies, the impact of different factors on the score is comprehensively considered, avoiding errors caused by a single factor, making the final regional defect score more comprehensive and reliable; under different acquisition conditions, by adjusting and correcting the score, the adaptability of the system is ensured, and performance fluctuations caused by changes in conditions are avoided; by accurately correcting the regional defect score, the ability to identify thread quality problems is improved, the quality control of prestressed anchor clips is enhanced, and it is ensured that it meets the high-precision requirements in the project; the multi-level correction mechanism effectively avoids the interaction between different error sources, reduces error accumulation, ensures high-precision performance of each step, and ultimately improves the detection effect of the overall system.

[0030] Furthermore, a coupling analysis is performed on all the corrected regional defect scores to obtain a comprehensive thread defect score of the prestressed anchor clip to be tested, and the score is compared with a preset thread defect threshold to generate a test result; Integrate the corrected defect scores of each thread inspection area; A coupled analysis method is used to comprehensively consider the mutual influence and weight distribution between defect scores in various regions. Different weight coefficients are assigned to different regions based on the differences in function and stress conditions of different parts of the thread. The starting and ending sections of the thread may have higher quality requirements due to processing difficulties or assembly requirements, so they can be assigned higher weights. The middle section may be assigned a relatively low but reasonable weight based on overall quality stability. A coupling algorithm is used to comprehensively calculate the defect scores of each area. The selection of the coupling algorithm should be based on the actual needs of thread quality assessment and the importance of each area to ensure that the comprehensive score can accurately reflect the quality status of the entire thread. Through the coupling algorithm, the defect scores of each area are weighted and summed to obtain the comprehensive thread defect score of the prestressed anchor clip to be tested; When the test results show that the comprehensive thread defect score of the prestressed anchor clip exceeds the set threshold, it indicates that there is a quality problem with the thread and further detailed analysis and correction are required. The comprehensive thread defect score is a comprehensive evaluation result of multiple thread inspection areas, including the quality assessment of the thread starting section, middle section, and end section. In each thread inspection area, deviation calculation is first performed based on the design requirements of the area and the actual collected geometric feature parameter set; specifically: For each thread inspection area, its geometric characteristic parameters are compared with the preset design standards; Calculate the deviation between the actual parameters and the design requirements and form a deviation vector; the deviation vector reflects the gap between the geometric characteristics of the area and the design requirements, including the deviation of the thread major diameter, pitch, tooth profile angle and tooth profile curvature; After obtaining the geometric feature deviation vector of each thread inspection area, serial adjustments are made based on the axial position relationship of the thread. The starting, middle, and ending segments of the thread differ in function and force, so their deviations are serially adjusted to ensure that the deviation vector of each area can be effectively analyzed and adjusted throughout the entire thread axis: By serializing and adjusting the geometric feature deviation vectors of all thread inspection areas, an internal hole thread feature deviation matrix is ultimately formed. This matrix integrates the deviation information of each inspection area and reflects the overall thread deviation. Each row of the deviation matrix represents the geometric feature deviation of an inspection area, and each column corresponds to a specific geometric feature. The internal thread feature deviation matrix is input into a preset defect identification and tracing model to identify defect types and adjust process parameters. The defect identification and tracing model can automatically identify different types of thread defects based on historical data and process experience. The defect identification and tracing model can also output defect types and corresponding process parameter adjustment strategies. Finally, by combining defect types with process parameter adjustment strategies, real-time feedback on the production process is generated; The internal thread feature deviation matrix is: Among them, d nm Indicates the deviation value between the mth geometric feature parameter of the nth thread detection area and the design requirements.

[0031] In this embodiment, the defect scores of each thread inspection area are integrated to comprehensively analyze the quality of different parts of the thread. The defect scores of different areas are weighted by the coupled analysis method to ensure that the scores are more consistent with the differences in function and stress of each area, thereby providing a more accurate quality assessment. The coupled algorithm accurately reflects the overall quality of the thread by comprehensively considering the mutual influence and weight distribution between the defect scores of each area. It also calculates the deviation of geometric feature parameters for each thread inspection area and forms a deviation vector, which helps to reveal the deviation between the actual thread geometry and the design requirements, and comprehensively reflects the quality fluctuations and defect types of the thread. By inputting the deviation matrix into the defect identification and traceability model, automatic identification of thread defect types and real-time adjustment of process parameters are achieved. The combination of defect identification and process parameter adjustment can provide immediate feedback for the production process, thereby optimizing the production process and reducing the incidence of thread defects. Comprehensive thread defect scoring and its combination with process adjustment strategies can help identify potential quality issues in advance and prevent substandard products from entering the market. This approach can improve the stability and reliability of the production process and ensure product quality.

[0032] like Figure 2 As shown, a visual inspection system for inner hole threads of a prestressed anchor clip of the present invention specifically includes the following modules: A region division module divides the internal thread of the prestressed anchor clip to be tested into structural regions along the axial direction to obtain at least three thread detection regions, wherein the thread detection regions include a thread starting section, a thread middle section, and a thread ending section; An image acquisition module is configured to acquire an in-hole image of each thread detection area, obtain a corresponding area image, and obtain a corresponding acquisition condition parameter set, wherein the acquisition condition parameter set includes light intensity, shooting angle, and equipment vibration intensity; a feature extraction module, performing geometric feature extraction on the regional image to obtain a geometric feature parameter set corresponding to each thread detection region, wherein the geometric feature parameter set includes thread major diameter, thread minor diameter, thread pitch, thread profile angle, and thread profile curvature; a defect feature evaluation module, for each of the thread inspection areas, inputting the geometric feature parameter set into a thread defect feature evaluation model corresponding to the area to obtain an area defect score; a scoring correction module, which corrects the regional defect score by taking into account the error influence of the acquisition condition parameter set on the actual feature; The comprehensive analysis module performs a coupling analysis on all the corrected regional defect scores to obtain a comprehensive thread defect score of the prestressed anchor clip to be tested, and compares it with a preset thread defect threshold to generate a test result.

[0033] The system comprehensively evaluates thread quality by extracting multiple geometric feature parameters rather than relying on a single feature parameter. This helps to more accurately reflect the actual quality of the thread and avoids other potential defects being masked by a single qualified feature. The system divides the internal thread of the prestressed anchor clip into multiple test areas along the axial direction and independently evaluates each area. This allows the system to highlight key quality issues in each area, as different areas have different functions and stress conditions, and therefore different quality requirements. The system records the acquisition condition parameter set during image acquisition and considers the error impact of these parameters on actual features in the scoring correction module. This helps improve the accuracy and reliability of defect scoring, as changes in acquisition conditions may affect image quality and feature extraction results. The system uses a comprehensive analysis module to couple all corrected regional defect scores to derive a comprehensive thread defect score for the prestressed anchor clip under test. This more comprehensively reflects the overall quality of the anchor clip's internal thread, providing a scientific basis for production decisions. The automated inspection method based on machine vision can quickly and accurately obtain thread images and geometric feature parameters, reducing the subjectivity and errors of manual inspection. At the same time, the system uses a preset thread defect feature evaluation model to quickly score and compare, improving inspection efficiency. In view of the miniaturization of threaded structures and the complexity of processing technology, the system can adapt to and accurately detect defects in various complex threaded structures through precise area division and feature extraction methods; it helps to ensure the processing quality of the inner hole thread of the prestressed anchor clip and improve the mechanical bite performance and long-term service reliability of the anchor system.

[0034] In addition, the present application also provides an electronic device, including a bus, a transceiver, a memory, a processor, and a computer program stored in the memory and runnable on the processor. The transceiver, the memory, and the processor are respectively connected via a bus. When the computer program is executed by the processor, each process of the above-mentioned method embodiment for controlling output data is implemented, and the same technical effect can be achieved. To avoid repetition, it will not be described here.

[0035] The above is only a preferred embodiment of the present invention. It should be pointed out that for ordinary technicians in this technical field, several improvements and modifications can be made without departing from the technical principles of the present invention. These improvements and modifications should also be regarded as the scope of protection of the present invention.

Claims

1. A method for visually inspecting the inner thread of a prestressed anchor clip, characterized in that: The method comprises: Divide the inner hole thread of the prestressed anchor clip to be tested into structural areas along the axial direction to obtain at least three thread detection areas; Performing in-hole image acquisition on each of the thread detection areas to obtain a corresponding area image and a corresponding acquisition condition parameter set; Performing geometric feature extraction on the area image to obtain a geometric feature parameter set corresponding to each thread detection area; For each thread inspection area, inputting the geometric feature parameter set into the thread defect feature evaluation model corresponding to the area to obtain a regional defect score; Taking into account the error influence of the acquisition condition parameter set on the actual features, the regional defect score is corrected; A coupling analysis is performed on all the corrected regional defect scores to obtain a comprehensive thread defect score of the prestressed anchor clip to be tested, which is then compared with a preset thread defect threshold to generate a test result.

2. The method for visually inspecting the inner thread of a prestressed anchor clip according to claim 1, characterized in that: In response to the combined thread defect score being higher than a preset thread defect threshold; For each of the thread detection areas, a deviation calculation is performed between the corresponding geometric feature parameter set and the design requirements of the inner hole thread of the prestressed anchor clip to be tested, so as to obtain a geometric feature deviation vector corresponding to the thread detection area; According to the positional relationship of the thread axis, all the geometric feature deviation vectors are serially adjusted to obtain the internal hole thread feature deviation matrix; The internal hole thread feature deviation matrix is input into a preset defect identification and tracing model, and the defect type and the process parameter adjustment strategy corresponding to the defect are output.

3. The method for visually inspecting the inner thread of a prestressed anchor clip according to claim 2, wherein: The internal thread feature deviation matrix is: Among them, d nm Indicates the deviation value between the mth geometric feature parameter of the nth thread detection area and the design requirements.

4. The method for visually inspecting the inner thread of a prestressed anchor clip according to claim 1, wherein: The thread detection area includes a thread starting section, a thread middle section and a thread ending section.

5. The method for visually inspecting the inner thread of a prestressed anchor clip according to claim 1, wherein: The acquisition condition parameter set includes light intensity, shooting angle and device vibration intensity.

6. The method for visually inspecting the inner thread of a prestressed anchor clip according to claim 1, wherein: The geometric feature parameter set includes thread major diameter, thread minor diameter, thread pitch, thread profile angle and thread profile curvature.

7. The method for visually inspecting the inner thread of a prestressed anchor clip according to claim 1, wherein: Taking into account the error influence of the acquisition condition parameter set on the actual features, the regional defect score is corrected, and the correction steps include illumination intensity correction, shooting angle correction, equipment vibration intensity correction, geometric feature extraction error correction and environmental factor correction.

8. A visual inspection system for inner hole threads of prestressed anchor clips, characterized in that: The system comprises: A region division module divides the internal thread of the prestressed anchor clip to be tested into structural regions along the axial direction to obtain at least three thread detection regions, wherein the thread detection regions include a thread starting section, a thread middle section, and a thread ending section; An image acquisition module is configured to acquire an in-hole image of each thread detection area, obtain a corresponding area image, and obtain a corresponding acquisition condition parameter set, wherein the acquisition condition parameter set includes light intensity, shooting angle, and equipment vibration intensity; a feature extraction module, performing geometric feature extraction on the regional image to obtain a geometric feature parameter set corresponding to each thread detection region, wherein the geometric feature parameter set includes thread major diameter, thread minor diameter, thread pitch, thread profile angle, and thread profile curvature; a defect feature evaluation module, for each of the thread inspection areas, inputting the geometric feature parameter set into a thread defect feature evaluation model corresponding to the area to obtain an area defect score; a scoring correction module, which corrects the regional defect score by taking into account the error influence of the acquisition condition parameter set on the actual feature; The comprehensive analysis module performs a coupling analysis on all the corrected regional defect scores to obtain a comprehensive thread defect score of the prestressed anchor clip to be tested, and compares it with a preset thread defect threshold to generate a test result.

9. An electronic device comprising a bus, a transceiver, a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein the transceiver, the memory, and the processor are connected via the bus, wherein: When the computer program is executed by the processor, the steps of the method according to any one of claims 1 to 7 are implemented.

10. A computer-readable storage medium having a computer program stored thereon, characterized in that: When the computer program is executed by a processor, the steps of the method according to any one of claims 1 to 7 are implemented.

Images