A heterogeneous soft wire serial automated quality detection method and device

Abstract

This invention discloses a serial automated quality inspection method and apparatus for heterogeneous flexible wires, belonging to the field of inspection. This application acquires production order data and controls wire processing equipment to complete processes such as feeding, wire cutting, heat shrink tubing assembly, terminal riveting, heat shrink tubing baking, and serial number etching to obtain the wire to be inspected. Then, a CCD vision system is used to acquire multi-dimensional image data of the wire. Combined with an improved gradient function, texture analysis method based on 3D surface deformation and optical reflection models, and feature matching algorithm, feature parameters of the wire surface, terminals, and heat shrink tubing are extracted to achieve pixel-level defect detection. Finally, the inspection results are decoded, the actual wire number and terminal shape are extracted, and compared with the production order data to determine whether the wire quality is compliant. This application effectively solves the problem of the inability to accurately and efficiently inspect the quality of heterogeneous flexible wires in existing technologies.

Description

Technical Field

[0001] This invention relates to the field of testing, and more particularly to a method and apparatus for automated serial quality testing of heterogeneous flexible wires. Background Technology

[0002] Heterogeneous flexible wires are widely used in industrial control equipment, automotive electronics, consumer electronics and other fields because they can be flexibly adapted in terms of wire diameter, terminal type and length. Their quality directly affects the electrical connection reliability and operational stability of downstream equipment. Therefore, quality inspection after processing is a key link in the production process.

[0003] The current quality inspection process for heterogeneous flexible wires has several technical limitations: Firstly, there is a lack of coordination between the pre-inspection processing steps and the inspection process. Existing wire processing equipment mostly operates in batches with fixed parameters, failing to dynamically adjust processing precision based on the differentiated needs of heterogeneous wires in production orders. This results in some wires developing potential defects prematurely due to processing deviations, increasing the burden of subsequent inspections. Secondly, existing inspection methods mainly rely on manual visual inspection or traditional machine vision inspection. Manual inspection depends on the operator's experience, which is not only inefficient but also prone to missing latent defects such as terminal tilting and minor scratches on the wire surface due to fatigue. Traditional machine vision inspection is mostly designed for flexible wires of the same specification and cannot adapt to the characteristics of heterogeneous wires with large differences in wire diameter and diverse terminal shapes, making it easy to misjudge features during inspection.

[0004] In the visual imaging and defect recognition stages, existing technologies have further revealed their shortcomings: the imaging parameters of traditional CCD vision systems are mostly fixed settings, and the exposure and gain are not adjusted according to the surface material and terminal reflective characteristics of heterogeneous wires. This often results in overexposure of the terminal area and blurring of the wire surface texture, leading to insufficient accuracy in subsequent feature extraction. At the same time, the defect recognition process lacks simulation analysis of the physical morphology of the wire surface. Defects are judged only by simple grayscale comparison, which cannot accurately distinguish the difference between surface scratches and normal texture, nor can it quantify the degree of deviation of terminal riveting offset, and cannot achieve pixel-level accurate defect positioning.

[0005] In addition, the compliance judgment logic of the existing testing process is imperfect. The test results can only determine whether the wire has defects, but are not linked to the preset specifications in the production order (such as heat shrink tubing wire number and terminal model). As a result, some wires, although they have no obvious defects, still flow into the downstream process due to incorrect wire number markings or terminal models that do not match the order, causing assembly failures.

[0006] These shortcomings prevent existing technologies from accurately and efficiently testing the quality of heterogeneous flexible wires. Summary of the Invention

[0007] This invention provides a serial automated quality inspection method and apparatus for heterogeneous flexible wires, to solve the problem that existing technologies cannot accurately and efficiently inspect the quality of heterogeneous flexible wires.

[0008] In a first aspect, this application provides a method for automated serial quality inspection of heterogeneous flexible wires, including: Obtain production order data; Based on the production order data, process control instructions are issued to the preset wire processing equipment, and the heterogeneous soft wire to be tested is obtained based on the process control instructions. The preset CCD vision system is controlled to align with the preset detection area of ​​the heterogeneous soft wire to be detected, and multi-dimensional image data of the wire to be detected is acquired. Based on the multi-dimensional image data, combined with the improved gradient function, the preset texture analysis method based on 3D surface deformation and optical reflection model and the preset feature matching algorithm, the feature parameters of the wire surface, terminals and heat shrink tubing in the multi-dimensional image data are extracted to obtain the pixel-level defect detection results of heterogeneous soft wires. The pixel-level defect detection results are decoded to extract the actual wire number and terminal shape of the wire to be inspected; the actual wire number and terminal shape are compared with the production order data to obtain the wire quality compliance judgment result.

[0009] This application first obtains production order data containing the processing sequence, length, specifications, terminal type, and heat shrink tubing number of heterogeneous flexible wires. This provides a precise basis for subsequent processing and inspection, addressing the differentiated characteristics of heterogeneous wires and avoiding the "one-size-fits-all" problem caused by the lack of order data support in traditional inspection. Second, based on this order data, control commands are issued to the processing equipment, ensuring that processes such as wire feeding, cutting, and riveting match specific order requirements. This reduces initial deviations caused by mismatches between processing parameters and the characteristics of heterogeneous wires, lowering the difficulty of defect identification in subsequent inspection stages from the source. Third, the application controls a CCD vision system to selectively acquire multi-dimensional images of the wires to be inspected, providing a comprehensive and clear image foundation for defect identification and avoiding the problems of local overexposure or texture blurring caused by fixed parameters in traditional vision imaging. Furthermore, it utilizes... The improved gradient function can accurately extract the edge features of terminals and heat shrink tubing. Combined with the texture analysis method of 3D surface deformation and optical reflection model, it can simulate the real defect morphology to accurately identify surface scratches. Then, the defect parameters are quantified by feature matching algorithm to achieve pixel-level defect detection, which solves the pain points of traditional manual inspection being prone to missed detection and traditional machine vision being prone to misjudgment. Finally, the actual wire number and terminal shape extracted by decoding are compared with the preset parameters of the order, which not only completes the defect judgment, but also realizes the specification compliance verification. This avoids focusing only on defects and ignoring downstream assembly failures caused by specification mismatch. In the end, a closed-loop solution covering "processing-inspection-compliance" is formed, which significantly improves the accuracy and efficiency of heterogeneous soft wire inspection, ensures that product quality is fully adapted to downstream needs, and effectively solves the problem that existing technologies cannot accurately and efficiently inspect the quality of heterogeneous soft wires.

[0010] Furthermore, the process control instructions include: instructions for wire feeding, wire cutting to order length, heat shrink tubing cutting and assembly, end terminal riveting, heat shrink tubing positioning and baking, number tube laser etching, pull-out force testing, wire continuity testing, and AOI inspection; the step of issuing process control instructions to preset wire processing equipment based on the production order data, and obtaining the heterogeneous soft wire to be inspected based on the process control instructions, specifically: Production order data is broken down according to a preset process sequence to generate corresponding process control instructions for each workstation; the process control instructions include the action parameters of each workstation. Based on the process control instructions, corresponding instructions are sequentially issued to the wire feeding station, wire cutting station, heat shrink tubing assembly station, terminal riveting station, heat shrink tubing positioning and baking station, number tube laser etching station, pull-out force testing station, wire continuity testing station and AOI inspection station. The AOI inspection station verifies the abnormal information of the wires and removes wires with abnormal information. The abnormal information includes abnormal wire stripping, abnormal terminals, exposed copper in the wire harness, incorrect terminals, incorrect number tube position, missing number tube, terminals that are too short, crooked terminals, and missing terminals. After each workstation completes its respective process, the processed wire is transferred to a pre-set testing station to obtain the heterogeneous soft wire to be tested.

[0011] This application achieves automated control of wire processing equipment by breaking down production order data into a preset process sequence and generating corresponding process control instructions for each workstation. This method not only improves production efficiency but also ensures the accuracy and consistency of the production process. Specifically, the process control instructions contain the action parameters of each workstation, enabling each workstation to execute corresponding actions according to precise instructions, thereby guaranteeing high quality and high efficiency in wire processing. By sequentially issuing instructions to each workstation, the continuity and automation of the entire production process are ensured, reducing manual intervention and potential errors. Finally, the processed wire is automatically transferred to a preset inspection station, forming heterogeneous soft wire to be inspected, providing a foundation for subsequent high-precision inspection. This process not only improves production efficiency but also reduces errors in the production process through automation and precise control, thereby improving product quality and consistency.

[0012] Furthermore, the preset CCD vision system is aligned with a preset detection area of ​​the heterogeneous flexible wire to be detected, and multi-dimensional image data of the wire to be detected is acquired, specifically as follows: A preset detection area is determined for the heterogeneous flexible wire to be tested, and the preset detection area covers the wire surface, the terminal riveting part and the heat shrink tubing marking part; After controlling the preset CCD vision system to move to a position that matches the preset detection area, the initial imaging sharpness is calculated using a sharpness evaluation method based on the axial edge of the wire. Based on the initial image sharpness, combined with image contrast, the proportion of overexposed areas and the proportion of underexposed areas, a comprehensive quality error is constructed. Based on the comprehensive quality error, the exposure and gain of the CCD vision system are adjusted by a fuzzy PID controller until the comprehensive quality error is less than a preset threshold, thereby obtaining imaging parameters that meet the requirements of the adaptive detection. Based on the imaging parameters and the preset CCD vision system, images are acquired in the preset detection area, and images of the wire surface, terminal riveting part and heat shrink tubing marking part are acquired in sequence to form multi-dimensional image data of the wire to be detected.

[0013] This application ensures comprehensive inspection by precisely locating the pre-defined inspection area of ​​the heterogeneous flexible wire to be inspected, covering the wire surface, terminal riveting areas, and heat shrink tubing markings. The CCD vision system is moved to the appropriate position, and the initial image sharpness is calculated based on the sharpness evaluation method of the wire's axial edge, making the image quality assessment targeted. A comprehensive quality error is constructed by combining image contrast, the proportion of overexposed areas, and the proportion of underexposed areas, providing a quantitative basis for optimizing imaging parameters. The exposure and gain of the CCD vision system are adjusted by a fuzzy PID controller until the comprehensive quality error is less than a preset threshold, achieving automatic optimization of imaging parameters and ensuring that the image quality meets inspection requirements. Finally, images are acquired from the pre-defined inspection area based on the optimized imaging parameters, forming multi-dimensional image data of the wire to be inspected. This provides high-quality data support for subsequent defect detection, improving the accuracy and reliability of the inspection. This process not only improves inspection efficiency but also reduces the impact of human factors on image quality through automation and precise control, ensuring the consistency and accuracy of the inspection results.

[0014] Furthermore, based on the multi-dimensional image data, combined with an improved gradient function, a preset texture analysis method based on a 3D surface deformation and optical reflection model, and a preset feature matching algorithm, feature parameters of the wire surface, terminals, and heat shrink tubing in the multi-dimensional image data are extracted to obtain pixel-level defect detection results for heterogeneous flexible wires, specifically: Based on the improved gradient function, the sharpness of the axial edge of the wire is calculated by combining the width, height and pixel value of each image in the multi-dimensional image data. Based on the clarity, extract the coordinates of the contour edge points of the terminal area and the boundary line of the heat shrink tubing area to obtain the terminal edge features and the heat shrink tubing boundary features. Based on a preset 3D surface deformation model and a preset adaptive data balancing method, the morphological characteristics of surface scratches are simulated in the wire surface region of multi-dimensional image data; according to the morphological characteristics and a preset optical reflection model, the reflection intensity of each pixel on the wire surface is calculated. The reflection intensity is compared with the preset reflection intensity of a normal wire surface to identify the texture abnormal area corresponding to the surface scratch, thus forming the texture abnormal feature of the wire surface. Using a preset feature matching algorithm, the terminal edge features are compared with preset terminal standard contour parameters, the heat shrink tubing boundary features are compared with preset heat shrink tubing standard boundary parameters, and the wire surface texture anomaly features are compared with preset wire surface standard reflection parameters. Based on the comparison results, the pixel-level location, type, and degree of deviation of the defects are determined, and pixel-level defect detection results of heterogeneous soft wires are obtained.

[0015] This application utilizes multi-dimensional image data and an improved gradient function to first calculate the sharpness of the axial edge of the wire, providing a foundation for subsequent feature extraction. Based on the sharpness information, the coordinates of the contour edge points of the terminal area and the boundary line of the heat shrink tubing area are accurately extracted, forming terminal edge features and heat shrink tubing boundary features, ensuring the accuracy of feature extraction. Furthermore, using a pre-defined 3D surface deformation model and an adaptive data balancing method, the morphological characteristics of surface scratches are simulated on the wire surface, enhancing the model's adaptability to different defect types. The reflection intensity of each pixel on the wire surface is calculated using an optical reflection model and compared with normal reflection intensity to identify texture abnormalities, forming texture abnormality features on the wire surface. Finally, through a feature matching algorithm, the extracted features are compared with pre-defined standard parameters to accurately determine the pixel-level location, type, and degree of deviation of defects, thereby obtaining pixel-level defect detection results for heterogeneous soft wires. This process not only improves the accuracy and reliability of defect detection but also reduces manual intervention through automated feature extraction and matching, improving detection efficiency.

[0016] Furthermore, the production order data includes the processing sequence, wire length, wire specifications, terminal models at both ends, and heat shrink tubing wire number of the heterogeneous flexible wire; the pixel-level defect detection results are decoded to extract the actual wire number and terminal shape of the wire to be inspected; the actual wire number and terminal shape are compared with the production order data to obtain the wire quality compliance judgment result, specifically: Decode the heat shrink tubing identification image in the pixel-level defect detection results to extract the actual wire number of the wire to be inspected; Morphological analysis is performed on the terminal images in the pixel-level defect detection results to extract the terminal morphology of the wire to be inspected; The extracted actual wire number is compared with the preset wire number in the production order data, and the extracted terminal shape is compared with the standard shape corresponding to the terminal model in the production order data. The compliance of the wire quality is determined based on the results of two comparisons. If the actual wire number is consistent with the preset wire number and the terminal shape conforms to the standard shape, the wire quality is deemed compliant; if either comparison result is inconsistent, the wire quality is deemed non-compliant.

[0017] This application achieves precise acquisition of key information about the wire by decoding the heat shrink tubing identification image in the pixel-level defect detection results, extracting the actual wire number of the wire to be inspected, and performing morphological analysis on the terminal image to extract the terminal shape. By comparing this extracted information with the standard shapes of preset wire numbers and terminal models in the production order data, it is possible to accurately determine whether the wire meets production requirements. This comparison method ensures precise control of wire quality, improving production efficiency and product quality consistency. Specifically, if the actual wire number matches the preset wire number and the terminal shape conforms to the standard shape, the wire is deemed compliant; conversely, if either comparison result is inconsistent, the wire is deemed non-compliant. This process not only improves the automation level of inspection but also reduces human error and enhances the quality control capabilities of the production process.

[0018] Furthermore, the sharpness of the axial edge of the wire is calculated based on the improved gradient function, combined with the width, height, and pixel values ​​of each image in the multi-dimensional image data, specifically as follows: Determine the target image corresponding to the terminal area and the heat shrink tubing area in the multi-dimensional image data; Select two pixels with a fixed pixel interval along the axial direction in the target image line by line, read the gray values ​​of the two pixels and calculate the absolute value of the difference between the gray values; The absolute values ​​of the grayscale differences across all rows of the target image are summed to obtain the total grayscale difference along the axial edges. Based on the preset effective width of the target image, and combined with the sum of the grayscale differences of the axial edges, the sharpness of the axial edges of the wire is calculated.

[0019] This application accurately locates key detection areas by identifying target images corresponding to the terminal region and heat shrink tubing region in multi-dimensional image data. It selects two pixels with a fixed pixel interval along the axial direction in the target image row by row, and calculates the absolute value of the difference in grayscale values ​​between these two pixels, thus quantifying the intensity of edge information. The absolute values ​​of the grayscale differences across all rows are summed to obtain the total grayscale difference along the axial edge. This operation integrates the image's edge features, providing comprehensive data support for sharpness calculation. Finally, combined with the preset effective width of the target image, the sharpness of the wire's axial edge is calculated, achieving a quantitative assessment of image quality. This process not only improves the accuracy of image processing but also provides reliable quality indicators for subsequent defect detection, ensuring the accuracy and reliability of the detection results.

[0020] Secondly, this application provides a serial automated quality inspection device for heterogeneous flexible wires. The serial automated quality inspection device for heterogeneous flexible wires includes: The first acquisition module is used to acquire production order data; The instruction issuing module is used to issue process control instructions to the preset wire processing equipment according to the production order data, and to obtain the heterogeneous soft wire to be tested based on the process control instructions; The second acquisition module is used to control the preset CCD vision system to align with the preset detection area of ​​the heterogeneous soft wire to be detected, and to acquire multi-dimensional image data of the wire to be detected. The feature extraction module is used to extract feature parameters of wire surface, terminals and heat shrink tubing in the multi-dimensional image data based on the multi-dimensional image data, combined with an improved gradient function, a preset texture analysis method based on 3D surface deformation and optical reflection model and a preset feature matching algorithm, so as to obtain pixel-level defect detection results of heterogeneous soft wires. The judgment module is used to decode the pixel-level defect detection results, extract the actual wire number and terminal shape of the wire to be inspected, and compare the actual wire number and terminal shape with the production order data to obtain the wire quality compliance judgment result.

[0021] The heterogeneous flexible wire serial automated quality inspection device of this application achieves fully automated inspection from production order data acquisition to wire quality judgment through the synergistic effect of various functional modules. First, the first acquisition module accurately acquires production order data including processing sequence, wire length, specifications, terminal type, and heat shrink tubing wire number, providing precise guidance for subsequent processes. Based on this data, the instruction issuing module precisely controls the wire processing equipment to complete each process from material loading to laser etching, ensuring that the processing meets order requirements and yields the heterogeneous flexible wire to be inspected. The second acquisition module controls the CCD vision system to align with the preset inspection area, acquiring multi-dimensional image data to provide comprehensive visual information for defect detection. The feature extraction module utilizes an improved gradient function, 3D surface deformation and optical reflection model, and feature matching algorithm to extract feature parameters of the wire surface, terminals, and heat shrink tubing from the multi-dimensional image data, achieving pixel-level defect detection and improving detection accuracy and reliability. The judgment module decodes the inspection results, extracts the actual wire number and terminal shape, and compares them with the production order data to determine whether the wire quality complies with regulations, achieving precise control over production quality. The entire device, through its modular design, improves testing efficiency, reduces manual intervention, enhances quality control capabilities in the production process, and ensures high-quality production of heterogeneous soft wires.

[0022] Furthermore, the preset CCD vision system is aligned with a preset detection area of ​​the heterogeneous flexible wire to be detected, and multi-dimensional image data of the wire to be detected is acquired, specifically as follows: A preset detection area is determined for the heterogeneous flexible wire to be tested, and the preset detection area covers the wire surface, the terminal riveting part and the heat shrink tubing marking part; After controlling the preset CCD vision system to move to a position that matches the preset detection area, the initial imaging sharpness is calculated using a sharpness evaluation method based on the axial edge of the wire. Based on the initial image sharpness, combined with image contrast, the proportion of overexposed areas and the proportion of underexposed areas, a comprehensive quality error is constructed. Based on the comprehensive quality error, the exposure and gain of the CCD vision system are adjusted by a fuzzy PID controller until the comprehensive quality error is less than a preset threshold, thereby obtaining imaging parameters that meet the requirements of the adaptive detection. Based on the imaging parameters and the preset CCD vision system, images are acquired in the preset detection area, and images of the wire surface, terminal riveting part and heat shrink tubing marking part are acquired in sequence to form multi-dimensional image data of the wire to be detected.

[0023] This application ensures comprehensive and targeted inspection by precisely defining the pre-defined inspection area of ​​the heterogeneous flexible wire to be inspected, covering the wire surface, terminal riveting areas, and heat shrink tubing markings. After controlling the CCD vision system to move to the appropriate position, the initial imaging sharpness is calculated using a sharpness evaluation method based on the axial edge of the wire, providing a foundation for subsequent image quality optimization. A comprehensive quality error is constructed by combining image contrast, the proportion of overexposed areas, and the proportion of underexposed areas; this quantitative indicator provides a scientific basis for adjusting imaging parameters. The exposure and gain of the CCD vision system are dynamically adjusted using a fuzzy PID controller until the comprehensive quality error is less than a preset threshold, achieving automatic optimization of imaging parameters and ensuring that the imaging quality meets inspection requirements. Finally, images are acquired from the pre-defined inspection area based on the optimized imaging parameters, forming multi-dimensional image data, providing high-quality data support for subsequent defect detection and quality assessment. This process not only improves the automation of image acquisition but also enhances image quality through precise control of imaging parameters, thereby improving the accuracy and reliability of defect detection and enhancing the performance and efficiency of the entire inspection system.

[0024] Furthermore, based on the multi-dimensional image data, combined with an improved gradient function, a preset texture analysis method based on a 3D surface deformation and optical reflection model, and a preset feature matching algorithm, feature parameters of the wire surface, terminals, and heat shrink tubing in the multi-dimensional image data are extracted to obtain pixel-level defect detection results for heterogeneous flexible wires, specifically: Based on the improved gradient function, the sharpness of the axial edge of the wire is calculated by combining the width, height and pixel value of each image in the multi-dimensional image data. Based on the clarity, extract the coordinates of the contour edge points of the terminal area and the boundary line of the heat shrink tubing area to obtain the terminal edge features and the heat shrink tubing boundary features. Based on a preset 3D surface deformation model and a preset adaptive data balancing method, the morphological characteristics of surface scratches are simulated in the wire surface region of multi-dimensional image data; according to the morphological characteristics and a preset optical reflection model, the reflection intensity of each pixel on the wire surface is calculated. The reflection intensity is compared with the preset reflection intensity of a normal wire surface to identify the texture abnormal area corresponding to the surface scratch, thus forming the texture abnormal feature of the wire surface. Using a preset feature matching algorithm, the terminal edge features are compared with preset terminal standard contour parameters, the heat shrink tubing boundary features are compared with preset heat shrink tubing standard boundary parameters, and the wire surface texture anomaly features are compared with preset wire surface standard reflection parameters. Based on the comparison results, the pixel-level location, type, and degree of deviation of the defects are determined, and pixel-level defect detection results of heterogeneous soft wires are obtained.

[0025] This application utilizes multi-dimensional image data and an improved gradient function to first calculate the sharpness of the axial edge of the wire, providing a foundation for subsequent feature extraction. Based on the sharpness information, the coordinates of the contour edge points of the terminal area and the boundary line of the heat shrink tubing area are accurately extracted, forming terminal edge features and heat shrink tubing boundary features, ensuring the accuracy of feature extraction. Furthermore, using a pre-defined 3D surface deformation model and an adaptive data balancing method, the morphological characteristics of surface scratches are simulated on the wire surface, enhancing the model's adaptability to different defect types. The reflection intensity of each pixel on the wire surface is calculated using an optical reflection model and compared with normal reflection intensity to identify texture abnormalities, forming texture abnormality features on the wire surface. Finally, through a feature matching algorithm, the extracted features are compared with pre-defined standard parameters to accurately determine the pixel-level location, type, and degree of deviation of defects, thereby obtaining pixel-level defect detection results for heterogeneous soft wires. This process not only improves the accuracy and reliability of defect detection but also reduces manual intervention through automated feature extraction and matching, improving detection efficiency.

[0026] Furthermore, the pixel-level defect detection results are decoded to extract the actual wire number and terminal shape of the wire to be inspected; the actual wire number and terminal shape are compared with the production order data to obtain the wire quality compliance judgment result, specifically: Decode the heat shrink tubing identification image in the pixel-level defect detection results to extract the actual wire number of the wire to be inspected; Morphological analysis is performed on the terminal images in the pixel-level defect detection results to extract the terminal morphology of the wire to be inspected; The extracted actual wire number is compared with the preset wire number in the production order data, and the extracted terminal shape is compared with the standard shape corresponding to the terminal model in the production order data. The compliance of the wire quality is determined based on the results of two comparisons. If the actual wire number is consistent with the preset wire number and the terminal shape conforms to the standard shape, the wire quality is deemed compliant; if either comparison result is inconsistent, the wire quality is deemed non-compliant.

[0027] This application achieves precise acquisition of key information about the wire by decoding the heat shrink tubing identification image in the pixel-level defect detection results, extracting the actual wire number of the wire to be inspected, and performing morphological analysis on the terminal image to extract the terminal shape. By comparing this extracted information with the standard shapes of preset wire numbers and terminal models in the production order data, it is possible to accurately determine whether the wire meets production requirements. This comparison method ensures precise control of wire quality, improving production efficiency and product quality consistency. Specifically, if the actual wire number matches the preset wire number and the terminal shape conforms to the standard shape, the wire is deemed compliant; conversely, if either comparison result is inconsistent, the wire is deemed non-compliant. This process not only improves the automation level of inspection but also reduces human error and enhances the quality control capabilities of the production process. Attached Figure Description

[0028] Figure 1 : A schematic flowchart of an embodiment of the serial automated quality inspection method for heterogeneous flexible wires provided in this application; Figure 2 : A schematic diagram of an embodiment of the production process provided in this application; Figure 3 : A schematic diagram of the structure of one embodiment for comparing image results provided in this application; Figure 4 This is a schematic diagram of an embodiment of the automated quality inspection device for heterogeneous flexible wires provided in this application. Detailed Implementation

[0029] The technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.

[0030] Example 1 Please refer to Figure 1 In order to solve the problem that existing technologies cannot accurately and efficiently detect the quality of heterogeneous flexible wires, this invention provides a serial automated quality detection method for heterogeneous flexible wires, including steps S01-S05.

[0031] S01: Obtain production order data; S02: Based on the production order data, issue process control instructions to the preset wire processing equipment, and obtain the heterogeneous soft wire to be tested based on the process control instructions; In a preferred embodiment of this invention, the step of issuing process control instructions to a preset wire processing equipment based on the production order data, and obtaining the heterogeneous soft wire to be tested based on the process control instructions, specifically involves: Production order data is broken down according to a preset process sequence to generate corresponding process control instructions for each workstation; the process control instructions include the action parameters of each workstation. Based on the process control instructions, corresponding instructions are sequentially issued to the wire feeding station, wire cutting station, heat shrink tubing assembly station, terminal riveting station, heat shrink tubing positioning and baking station, number tube laser etching station, pull-out force testing station, wire continuity testing station and AOI inspection station. After each workstation completes its respective process, the processed wire is transferred to the preset testing station to obtain the heterogeneous soft wire to be tested. The production process flow of this application is as follows: Figure 2 As shown, Figure 2 The process of processing and quality inspection of heterogeneous flexible wires was demonstrated: First, the 1.58 and 2.5 specification wires were loaded and cut. Then, the number tubes were installed, the number tubes were injected, the terminals were loaded and riveted, the 2.5 terminal was loaded and riveted, and the heat-shrink number tubes were applied. Next, the wires that passed the continuity test and the pull-out force test entered the AOI inspection stage. If the AOI inspection failed, the defective wires were eliminated. After the wires were connected and transferred, they were sent to the wire storage production line. The method to ensure that each workstation completes each process sequentially is as follows: The production information system sends production order information to the wire processing equipment. The equipment then breaks down the production order step-by-step according to the process flow chart. Each production order contains complete information about the wire to be processed, including the processing sequence, length, specifications, terminal types at both ends, and heat shrink tubing wire number. This production order information is then sequentially sent to each workstation for execution. a. Wire feeding: The entire roll of wire is fed into the wire inlet of the feeding section, the positioning block is pressed down, and the wire feeding is completed. When there is no wire material, the photosensitive material sensor will automatically activate the system alarm. The system will promptly remind the staff to replenish the material through indicator lights, panel display, and buzzer. If replenishment is not completed in time, when the equipment reaches the tail end of the wire production, the system will automatically stop and remind the staff to replenish the material again through indicator lights, panel display, and buzzer. b. Loading heat shrink tubing: The entire roll of heat shrink tubing is placed on a dedicated heat shrink tubing bracket, the wire end is fed to the servo feed port, the positioning block is pressed down, and the heat shrink tubing installation is completed; c. Wire feeding and cutting: After receiving the wire processing information from the production order, the wire feeding and cutting station initiates automatic wire feeding at 1.5mm² (or 2.5mm²) lengths, according to the requirements. After initial feeding, the wire clamping component picks up the wire head, while the automatic feeding system simultaneously feeds the wire according to the length specified in the production order. Once feeding is complete, another clamping component at the other end picks up the wire tail, and the automatic feeding section cuts it off, completing the wire removal process. d. Assemble the heat shrink tubing: The wire is automatically moved to the heat shrink tubing installation point, and both ends are simultaneously entered into the tubing connection interface. The heat shrink tubing equipment is started to cut the heat shrink tubing and install it into the corresponding ends of the wire. The heat shrink tubing installation process is then completed. e. Crimped terminals: The crimping terminal station receives the wire terminal model information from the production order. The wire, already fitted with heat-shrink tubing, is moved to the terminal crimping and fitting equipment. One end of the wire is inserted into the corresponding terminal crimping interface, and the crimping machine is started to crimp the terminal. After crimping, the other end is inserted into the corresponding terminal crimping interface, and the crimping machine is started again, completing the crimping process. f. Heat shrink tubing positioning and baking: The linear actuator moves the pre-riveted wire to the hot-melt process, simultaneously advancing the component. The hot-melt component cylinder lifts the component to perform sleeve hot-melt. The set time is up when hot-melt is complete. The hot-melt component then descends. g. Laser heat shrink tubing: The laser marking equipment at the laser station receives the wire number information from the production order. It controls a linear actuator to transfer the wire, whose ends have been heat-fused, to the laser station. The wire clamp positions the wire head. After positioning, it begins laser marking the wire number and one-dimensional barcode on the left side of the wire. After completion, it laser marks the wire number and one-dimensional barcode on the right side. Once laser marking is complete, the positioning clamp is released. Specific laser settings are shown in Table 1 below. Table 1 Laser parameter settings h. Pull-out force test: This process is not used during normal production; one wire is randomly selected for pull-out force testing every 100 wires produced. After the wires are riveted with terminals, they are directly sent to this process for pull-out force testing. A pull-out force tester is used until the terminals at both ends detach, and the tester transmits the data to the MES system. i. Wire continuity test: The wire with both ends riveted is transferred to the interface of the functional testing equipment. After the continuity testing device clamps both ends of the wire, the testing instrument performs continuity testing on the wire. When the testing instrument determines the result is OK, the wire is sent to the AOI image testing equipment to continue the next process of production. If the testing instrument determines the result is NG, the relevant information is uploaded to the MES system for defect statistics and displayed on the system display screen. The mover sends this defective product to the defective material pick-up point, and the defective material pick-up mechanism places the defective product into the defective product frame. The equipment then picks up the wire again to continue production. j. Automated Optical Inspection (AOI) of Wires: The AOI (Automated Optical Inspection) equipment receives wire manufacturing information from the production order, including the terminal models at both ends of the wire and the wire number of the heat shrink tubing. Wires that pass continuity testing are then subjected to automated optical inspection. The imaging system performs visual inspection on the wire's appearance and crimping shape, and verifies the alphanumeric strings and barcodes printed on the heat shrink tubing against the relevant information in the production order. If the AOI result is OK, the linear actuator delivers the OK wire to the variable-pitch pick-up area. If the AOI result is NG (Not Acceptable), the system uploads the relevant information to the MES (Manufacturing Execution System) and performs defect statistics, displaying the data on the system screen. The linear actuator delivers the defective product to the defective pick-up area, where the defective pick-up mechanism places the defective product in a defective product frame, and the equipment retrieves the wire again to continue production. Figure 3 As shown, Figure 3 This is a schematic diagram comparing the image results of this application. It shows the comparison between various defective wire harnesses and normal wire harnesses (1X2-24 specification wire harnesses marked "normal wire harnesses" below). The defects include exposed wires, soldered terminals, exposed copper in the wire harness, incorrect terminals, incorrect number tube position, missing number tube, terminals that are too short, misaligned terminals, missing terminals, etc. It can intuitively demonstrate the ability of image inspection to identify and distinguish various wire harness defects.

[0032] k. Wire circulation: The wires that pass the AOI inspection are moved to the docking carrier, and then the wires are moved to the wire assembly line frame. At this point, the entire wire processing procedure is completed.

[0033] S03: Control the preset CCD vision system to align with the preset detection area of ​​the heterogeneous soft wire to be detected, and acquire multi-dimensional image data of the wire to be detected; In a preferred embodiment of this invention, the preset CCD vision system is aligned with a preset detection area of ​​the heterogeneous flexible wire to be detected, and multi-dimensional image data of the wire to be detected is acquired, specifically as follows: In this embodiment, to address the quality inspection requirements of heterogeneous flexible wires, the pre-defined inspection area for the wire to be inspected is first defined. This area must cover the key quality control points of the wire, specifically including the wire surface, terminal riveting areas, and heat shrink tubing markings. The wire surface is a high-risk area for defects such as scratches and dents; the terminal riveting areas directly affect the reliability of electrical connections; and the heat shrink tubing markings are related to specification compliance determination. By covering these three types of areas, blind spots in quality inspection can be avoided, ensuring comprehensive inspection. Subsequently, the preset CCD vision system is controlled to move towards the preset detection area to complete the position calibration: according to the distribution characteristics of the detection area, the spatial position and lens angle of the CCD camera are adjusted to ensure that the camera's field of view can fully cover the entire view of the terminal riveting part, the full length range of the wire surface, and the character area of ​​the heat shrink tubing marking part, while avoiding local occlusion and reflection dead angles caused by equipment structure or ambient light, ensuring that subsequent imaging can clearly present the details of each area. After the position calibration is completed, the initial imaging sharpness is calculated using a sharpness evaluation method based on the axial edge of the wire: For the images of the terminal riveting part and the heat shrink tubing marking part, two pixels with a fixed pixel interval in the axial direction are selected row by row, the gray value of the two pixels is read and the absolute value of the gray value difference is calculated, the absolute values ​​of the gray value differences of all rows are accumulated and then divided by the product of the image height and the effective width (total image width minus the above interval number of pixels) to obtain the initial imaging sharpness, thereby determining whether the current imaging meets the edge recognition requirements; Based on the initial imaging clarity, a comprehensive quality error is constructed by further combining image contrast, overexposed area ratio, and underexposed area ratio. Image contrast is calculated by the grayscale difference within the region of interest (ROI) on the wire surface, reflecting the discernibility of surface texture details. Overexposed area ratio is the proportion of pixels in the image with excessively high grayscale values ​​that cause loss of detail, while underexposed area ratio is the proportion of pixels with excessively low grayscale values ​​that cause blurred details. By assigning appropriate weights to each indicator and performing a weighted summation, the comprehensive quality error is obtained to comprehensively evaluate whether the imaging quality meets the detection requirements. Based on the initial image sharpness described above, a comprehensive quality error is further constructed by combining image contrast, the proportion of overexposed areas, and the proportion of underexposed areas. Specifically: Sharpness Metric (S): Employs a modified Brenner gradient function, emphasizing the sharpness of the wire's axial (movement direction) edges. S = Σ_{y=1}^{H} Σ_{x=1}^{W-2} |I(x+2,y) - I(x,y)| / (H*(W-2)) where I(x,y) is the pixel grayscale value, and W and H are the image width and height. This formula enhances axial edge response when the wire moves along the x-axis.

[0034] Contrast Metric (C): Local contrast ratio of the ROI region (wire surface). C = (μ_high - μ_low) / (μ_high + μ_low + ε) where μ_high and μ_low are the mean values ​​of the top 5% and bottom 5% quantiles of the grayscale histogram within the ROI, and ε is the decimal point.

[0035] Percentage of overexposed / underexposed areas (O, U): O = Σ [I(x,y)>T_high] / N_pixels, U = Σ [I(x,y) <T_low] / N_pixels In the formula, the threshold T_high=250, T_low=10.

[0036] Feedback control algorithm (fuzzy PID controller), defining the overall quality error E: E = w1*(1 - S / S_target) + w2*(1 - C / C_target) + w3*O + w4*U; where the weights w1-w4 are adjustable, and S_target and C_target are preset target values.

[0037] ΔExposure = Kp * dE + Ki * ∫E dt + Kd * dE / dt; ΔGain = α * ΔExposure (α is the gain-exposure coupling coefficient); Kp, Ki, and Kd are tuned online using fuzzy rules: for example, when E is large and dE / dt is positive, Kp is increased.

[0038] The overall quality error is input into the fuzzy PID controller for parameter adjustment: if the overall quality error exceeds the preset threshold, it indicates that the current image clarity, contrast, or brightness does not meet the requirements. The fuzzy PID controller will dynamically adjust the exposure and gain of the CCD vision system according to the magnitude and trend of the error. When the error is large, a larger parameter adjustment is used to quickly optimize the image quality; when the error gradually decreases and approaches the threshold, the adjustment amplitude is reduced to avoid parameter oscillation, until the overall quality error is less than the threshold. At this time, the corresponding exposure and gain are the imaging parameters that meet the detection requirements. Finally, based on the determined imaging parameters, the CCD vision system is controlled to acquire images of the preset detection area: first, images of the terminal riveting area are acquired to ensure a clear view of the connection outline between the terminal and the wire; then, images of the wire surface are acquired, capturing images along the entire length of the wire to cover the surface conditions at different locations; finally, images of the heat shrink tubing markings are acquired, focusing on the marking character area to ensure clear character details. All acquired images are integrated in the logical order of "terminal riveting - wire surface - heat shrink tubing markings" to form multi-dimensional image data of the wire to be inspected, providing a complete and reliable image foundation for subsequent extraction of feature parameters of the wire surface, terminals, and heat shrink tubing.

[0039] S04: Based on the multi-dimensional image data, combined with the improved gradient function, the preset texture analysis method based on 3D surface deformation and optical reflection model and the preset feature matching algorithm, the feature parameters of the wire surface, terminals and heat shrink tubing in the multi-dimensional image data are extracted to obtain the pixel-level defect detection results of heterogeneous soft wires. In a preferred embodiment of this invention, based on the multi-dimensional image data, and combined with an improved gradient function, a preset texture analysis method based on a 3D surface deformation and optical reflection model, and a preset feature matching algorithm, feature parameters of the wire surface, terminals, and heat shrink tubing in the multi-dimensional image data are extracted to obtain pixel-level defect detection results for heterogeneous flexible wires. Specifically: In this embodiment, for feature extraction and defect detection of heterogeneous flexible wires, the sharpness of the axial edge of the wire is first calculated based on an improved gradient function: images corresponding to the terminal area and the heat shrink tubing area are selected from multi-dimensional image data. Combining the width, height and grayscale value of each pixel in these images, two pixels with a fixed pixel interval are selected row by row along the wire axis. The grayscale values ​​of the two points are read and the absolute value of the grayscale difference is calculated. The absolute values ​​of the grayscale differences of all rows are accumulated and divided by the product of the image height and the effective width (total image width minus the number of interval pixels mentioned above) to obtain the sharpness of the axial edge of the wire. This sharpness value directly reflects the identifiability of the terminal and heat shrink tubing edges, providing a basis for judgment in subsequent edge extraction.

[0040] Based on the sharpness calculated above, the edge features of the terminal and heat shrink tubing are further extracted: a sharpness threshold is set, and pixels with sharpness higher than the threshold are selected as edge candidate points (pixels with sharpness lower than the threshold are judged as non-edge areas and excluded); axial connectivity analysis is performed on the edge candidate points, retaining candidate points continuously distributed along the wire axis and eliminating isolated discrete points, and finally determining the coordinates of the outline edge points of the terminal area (precisely marking the position of each key pixel of the terminal outline) and the boundary line of the heat shrink tubing area (clarifying the boundary pixel trajectory between the heat shrink tubing and the wire body), thus forming the terminal edge features and the heat shrink tubing boundary features.

[0041] For the analysis of the texture features of the wire surface, the morphological features of surface scratches are first simulated based on a preset 3D surface deformation model and a preset adaptive data balancing method. Based on the defect-free area image of the wire surface in multi-dimensional image data, scratch deformation is superimposed on the height distribution of the defect-free surface through the 3D surface deformation model (the deformation morphology is set according to the depth and length characteristics of common scratches). At the same time, the adaptive data balancing method will refer to the recognition of various types of scratches in historical inspections. If the recognition rate of a certain type of scratch (such as shallow fine scratches) is low, the generation ratio of this type of scratch in the simulation process will be increased to ensure that the simulated samples can cover different types of surface defects and avoid recognition bias caused by uneven sample distribution. Based on this, the reflection intensity of each pixel on the wire surface is calculated using a pre-set optical reflection model. This optical reflection model calculates the reflection intensity by superimposing the diffuse reflection component and the specular reflection component. The diffuse reflection component is determined by the angle between the surface normal vector and the light source direction (the angle of the normal vector changes with different surface morphologies). The specular reflection component is determined by the angle between the reflection direction and the viewing direction, as well as the surface roughness. For the simulated scratch area, due to the change in surface morphology, the reflection intensity of the transition zone between the normal vector and the roughness will be significantly different from that of the defect-free area. The model can accurately calculate the reflection intensity value corresponding to this difference. The preset 3D surface deformation model is specifically as follows: 3D Surface Deformation Function: Superimposing scratch deformation onto the height map H(x,y) of a defect-free wire surface: H'(x,y) = H(x,y) + A * exp( -[(x-x0)² / (2σx²) + (y-y0)² / (2σy²)] ) In the formula, A is the scratch depth (negative value), (x0,y0) is the center point, and σx and σy are the attenuation in the length and width directions of the scratch.

[0042] The surface reflectivity of the wire is determined using a micro-facet model, with pixel reflection intensity I as follows: I = k_d * (N·L) + k_s * (R·V)^n / (N·V + ε) In the formula, k_d and k_s are the diffuse / specular reflection coefficients, respectively; N is the normal vector; L is the light source direction; V is the line of sight direction; R is the reflection direction; N at the scratch is calculated from the deformed H'(x,y), which significantly changes the specular shape; The calculated reflection intensity of each pixel on the wire surface is compared with the preset reflection intensity of a normal wire surface (a standard reflection intensity range obtained statistically from a large number of defect-free wire samples). If the difference between the actual reflection intensity of a pixel and the normal reflection intensity exceeds the preset deviation threshold, the area where the pixel is located is determined to be a texture abnormal area. All texture abnormal areas are integrated, and the pixel distribution range of the abnormal area is marked to form the texture abnormal feature of the wire surface. This feature directly corresponds to the location and range of scratches on the wire surface.

[0043] Then, a preset feature matching algorithm is used to compare each feature with the standard parameters: the coordinates of the contour edge points in the terminal edge features are compared with the preset standard contour parameters of the terminal (the standard pixel contour data corresponding to the terminal model in the order), and the positional deviation between the edge point coordinates and the standard contour is calculated; the boundary line in the heat shrink tubing boundary features is compared with the preset standard boundary parameters of the heat shrink tubing (the standard pixel trajectory of the position where the heat shrink tubing should be) to determine whether the boundary line is offset; the abnormal features of the wire surface texture are compared with the preset standard reflection parameters of the wire surface (the allowable range of reflection intensity deviation of a defect-free surface) to confirm whether the abnormal area conforms to the reflection features of a scratch defect. Defect generation probability is dynamically adjusted: P_i^{(t+1)} = P_i^{(t)} * (1 - β) + β * (1 - Recall_i^{(t)}) In the formula, P_i is the generation probability of the i-th type of defect; Recall_i^{(t)}: the recall rate of the current model for the i-th type of defect on the validation set; β is the smoothing coefficient, which drives the model to focus on weak defects; Based on the above three comparison results, the pixel-level defect detection results of heterogeneous flexible wires were finally determined: combining the deviation of the terminal edge point coordinates from the standard contour, the pixel-level position of the terminal edge offset (e.g., an edge point deviates from the standard position by 5 pixels) and the degree of deviation (the larger the deviation value, the more severe the offset); based on the comparison of the heat shrink tubing boundary line with the standard boundary, it was determined whether the heat shrink tubing has boundary anomalies (e.g., overall boundary line offset or local deformation) and abnormal positions; by comparing the surface texture anomalies with the standard reflection parameters, the pixel-level distribution range and severity of scratch defects were clarified (the larger the reflection intensity deviation, the deeper the scratch); integrating this information, a complete pixel-level defect detection result was formed, including defect pixel coordinates, defect type (terminal edge offset, heat shrink tubing boundary anomaly, surface scratch), and degree of deviation.

[0044] S05: Decode the pixel-level defect detection results to extract the actual wire number and terminal shape of the wire to be inspected; compare the actual wire number and terminal shape with the production order data to obtain the wire quality compliance judgment result.

[0045] In a preferred embodiment of this invention, the step of decoding the pixel-level defect detection result to extract the actual wire number and terminal shape of the wire to be inspected, and comparing the actual wire number and terminal shape with the production order data to obtain the wire quality compliance judgment result, specifically involves: In this embodiment, the pixel-level defect detection results of the aforementioned heterogeneous flexible wires are further processed to complete the quality compliance determination. The specific process is as follows: First, the image of the heat shrink tubing markings in the pixel-level defect detection results is decoded to extract the actual wire number of the wire to be inspected. This process uses a character recognition algorithm to identify the marking characters on the heat shrink tubing: the image is preprocessed (including noise reduction and enhancement of character edge contrast), then the character region is located by contour extraction, and subsequently, the character recognition model is called to parse the character sequence within the region to obtain the actual wire number containing letters, numbers, or symbols (e.g., "ECU-001"). If the marking characters are partially blurred or incompletely etched, the algorithm will combine contextual character features for completion verification to ensure the accuracy of the actual wire number extraction.

[0046] Simultaneously, images of the terminal riveting area are extracted from the pixel-level defect detection results, and the terminal morphology is analyzed to obtain key feature parameters. Specifically, the overall outline of the terminal is delineated using a contour extraction algorithm, the edge feature points of the terminal (such as the vertex of the terminal head and the transition point connecting with the wire) are identified, the contour dimensions of the terminal (such as length and width) and the riveting angle relative to the wire (the angle between the terminal axis and the wire axis) are calculated, and these parameters are integrated into the terminal morphology features of the wire to be inspected.

[0047] After extracting the actual wire number and terminal shape, both are compared with the preset information in the production order data. Specifically, the actual wire number is compared with the preset wire number in the production order data at the character level to check if the character sequence matches completely (including uppercase and lowercase letters, number order, and symbol position). The terminal shape features are compared with the standard shape parameters of the corresponding terminal model in the production order data (these parameters are pre-stored based on the design specifications of that terminal model, including the standard outline size range and allowable riveting angle deviation). The difference between the actual size and the standard size, and the deviation between the actual riveting angle and the standard angle are calculated to determine if they are within the preset allowable range.

[0048] Based on the above two comparison results, the final quality compliance of the wire is determined as follows: if the actual wire number is completely consistent with the preset wire number, and the deviations of various parameters of the terminal shape characteristics are all within the standard allowable range, then the heterogeneous flexible wire is judged to be compliant; if there are character differences between the actual wire number and the preset wire number (such as misaligned numbers or missing symbols), or the deviations of the terminal shape parameters exceed the allowable range (such as excessively large outline dimensions or excessively skewed riveting angles), then it is judged to be non-compliant, and the specific information of the non-compliant items (such as incorrect wire number location and terminal deviation value) is recorded to provide a basis for subsequent rework or screening.

[0049] This embodiment first acquires production order data containing the processing sequence, length, specifications, terminal type, and heat shrink tubing number of heterogeneous flexible wires. This provides a precise basis for subsequent processing and inspection, addressing the differentiated characteristics of heterogeneous wires and avoiding the "one-size-fits-all" problem caused by the lack of order data support in traditional inspection. Second, based on this order data, control commands are issued to the processing equipment, ensuring that processes such as wire feeding, cutting, and riveting match specific order requirements. This reduces initial deviations caused by mismatches between processing parameters and the characteristics of heterogeneous wires, lowering the difficulty of defect identification in subsequent inspection stages from the source. Third, the CCD vision system is controlled to selectively acquire multi-dimensional images of the wires to be inspected, providing a comprehensive and clear image foundation for defect identification and avoiding the problems of local overexposure or texture blurring caused by fixed parameters in traditional visual imaging. Furthermore, by… The improved gradient function can accurately extract the edge features of terminals and heat shrink tubing. Combined with the texture analysis method of 3D surface deformation and optical reflection model, it can simulate the real defect morphology to accurately identify surface scratches. Then, the defect parameters are quantified by feature matching algorithm to achieve pixel-level defect detection, which solves the pain points of traditional manual inspection being prone to missed detection and traditional machine vision being prone to misjudgment. Finally, the actual wire number and terminal shape extracted by decoding are compared with the preset parameters of the order, which not only completes the defect judgment, but also realizes the specification compliance verification. This avoids focusing only on defects and ignoring downstream assembly failures caused by specification mismatch. In the end, a closed-loop solution covering "processing-inspection-compliance" is formed, which significantly improves the accuracy and efficiency of heterogeneous soft wire inspection, ensures that product quality is fully adapted to downstream needs, and effectively solves the problem that existing technologies cannot accurately and efficiently inspect the quality of heterogeneous soft wires.

[0050] Example 2 Please refer to Figure 4 This is a serial automated quality inspection device for heterogeneous flexible wires provided in the embodiments of this application.

[0051] In this embodiment, the heterogeneous flexible wire serial automated quality inspection device includes a first acquisition module 10, an instruction issuing module 20, a second acquisition module 30, a feature extraction module 40, and a judgment module 50.

[0052] The first acquisition module 10 is used to acquire production order data; The instruction issuing module 20 is used to issue process control instructions to the preset wire processing equipment according to the production order data, and to obtain the heterogeneous soft wire to be tested based on the process control instructions; The second acquisition module 30 is used to control the preset CCD vision system to align with the preset detection area of ​​the heterogeneous soft wire to be detected, and acquire multi-dimensional image data of the wire to be detected. The feature extraction module 40 is used to extract feature parameters of the wire surface, terminals and heat shrink tubing in the multi-dimensional image data based on the multi-dimensional image data, combined with an improved gradient function, a preset texture analysis method based on 3D surface deformation and optical reflection model and a preset feature matching algorithm, so as to obtain pixel-level defect detection results of heterogeneous soft wires. The judgment module 50 is used to decode the pixel-level defect detection results, extract the actual wire number and terminal shape of the wire to be inspected, and compare the actual wire number and terminal shape with the production order data to obtain the wire quality compliance judgment result.

[0053] For ease of description and brevity, the embodiments of the device of the present invention include all the implementation methods in the above embodiments of the serial automated quality inspection method for heterogeneous flexible wires, and will not be repeated here.

[0054] The specific embodiments described above further illustrate the purpose, technical solution, and beneficial effects of the present invention. It should be understood that the above descriptions are merely specific embodiments of the present invention and are not intended to limit the scope of protection of the present invention. In particular, it should be noted that any modifications, equivalent substitutions, improvements, etc., made within the spirit and principles of the present invention should be included within the scope of protection of the present invention for those skilled in the art.

Claims

1. A serial automated quality inspection method for heterogeneous flexible wires, characterized in that, include: Obtain production order data; Based on the production order data, process control instructions are issued to the preset wire processing equipment, and the heterogeneous soft wire to be tested is obtained based on the process control instructions. The preset CCD vision system is controlled to align with the preset detection area of ​​the heterogeneous flexible wire to be detected, and multi-dimensional image data of the wire to be detected is acquired, specifically: A preset detection area is determined for the heterogeneous flexible wire to be tested, and the preset detection area covers the wire surface, the terminal riveting part and the heat shrink tubing marking part; After controlling the preset CCD vision system to move to a position that matches the preset detection area, the initial imaging sharpness is calculated using a sharpness evaluation method based on the axial edge of the wire. Based on the initial image sharpness, combined with image contrast, the proportion of overexposed areas and the proportion of underexposed areas, a comprehensive quality error is constructed. Based on the comprehensive quality error, the exposure and gain of the CCD vision system are adjusted by a fuzzy PID controller until the comprehensive quality error is less than a preset threshold, thereby obtaining imaging parameters that meet the requirements of the adaptive detection. Based on the imaging parameters and the preset CCD vision system, images are acquired in the preset detection area, and images of the wire surface, terminal riveting part and heat shrink tubing marking part are acquired in sequence to form multi-dimensional image data of the wire to be detected. Based on the multi-dimensional image data, combined with an improved gradient function, a preset texture analysis method based on 3D surface deformation and optical reflection models, and a preset feature matching algorithm, feature parameters of the wire surface, terminals, and heat shrink tubing in the multi-dimensional image data are extracted to obtain pixel-level defect detection results for heterogeneous flexible wires, specifically: Based on the improved gradient function, the sharpness of the axial edge of the wire is calculated by combining the width, height and pixel value of each image in the multi-dimensional image data. Based on the clarity, extract the coordinates of the contour edge points of the terminal area and the boundary line of the heat shrink tubing area to obtain the terminal edge features and the heat shrink tubing boundary features. Based on a preset 3D surface deformation model and combined with a preset adaptive data balancing method, the morphological characteristics of surface scratches are simulated in the wire surface region of multi-dimensional image data. Based on the morphological characteristics and combined with a preset optical reflection model, the reflection intensity of each pixel on the wire surface is calculated. The reflection intensity is compared with the preset reflection intensity of a normal wire surface to identify the texture abnormal area corresponding to the surface scratch, thus forming the texture abnormal feature of the wire surface. Using a preset feature matching algorithm, the terminal edge features are compared with preset terminal standard contour parameters, the heat shrink tubing boundary features are compared with preset heat shrink tubing standard boundary parameters, and the wire surface texture anomaly features are compared with preset wire surface standard reflection parameters. Based on the comparison results, the pixel-level location, defect type, and deviation degree of the defect are determined, and the pixel-level defect detection results of the heterogeneous soft wire are obtained. The pixel-level defect detection results are decoded to extract the actual wire number and terminal shape of the wire to be detected; The actual wire number and terminal shape are compared with the production order data to obtain the wire quality compliance judgment result.

2. The automated serial quality inspection method for heterogeneous flexible wires according to claim 1, characterized in that, The process control instructions include: wire feeding, wire cutting to order length, heat shrink tubing cutting and assembly, end terminal riveting, heat shrink tubing positioning and baking, number tube laser etching, pull-out force testing, wire continuity testing, and AOI inspection instructions; the process control instructions are issued to the preset wire processing equipment based on the production order data, and the heterogeneous soft wire to be inspected is obtained based on the process control instructions, specifically: Production order data is broken down according to a preset process sequence to generate corresponding process control instructions for each workstation; the process control instructions include the action parameters of each workstation. Based on the process control instructions, corresponding instructions are sequentially issued to the wire feeding station, wire cutting station, heat shrink tubing assembly station, terminal riveting station, heat shrink tubing positioning and baking station, number tube laser etching station, pull-out force testing station, wire continuity testing station and AOI inspection station. The AOI inspection station verifies the abnormal information of the wires and removes wires with abnormal information. The abnormal information includes abnormal wire stripping, abnormal terminals, exposed copper in the wire harness, incorrect terminals, incorrect number tube position, missing number tube, terminals that are too short, crooked terminals, and missing terminals. After each workstation completes its respective process, the processed wire is transferred to a pre-set testing station to obtain the heterogeneous soft wire to be tested.

3. The automated serial quality inspection method for heterogeneous flexible wires according to claim 1, characterized in that, The production order data includes the processing sequence, wire length, wire specifications, terminal models at both ends, and heat shrink tubing number of the heterogeneous flexible wires; the pixel-level defect detection results are decoded to extract the actual wire number and terminal shape of the wire to be inspected; the actual wire number and terminal shape are compared with the production order data to obtain the wire quality compliance judgment result, specifically: Decode the heat shrink tubing identification image in the pixel-level defect detection results to extract the actual wire number of the wire to be inspected; Morphological analysis is performed on the terminal images in the pixel-level defect detection results to extract the terminal morphology of the wire to be inspected; The extracted actual wire number is compared with the preset wire number in the production order data, and the extracted terminal shape is compared with the standard shape corresponding to the terminal model in the production order data. The compliance of the wire quality is determined based on the results of two comparisons. If the actual wire number is consistent with the preset wire number and the terminal shape conforms to the standard shape, the wire quality is deemed compliant; if either comparison result is inconsistent, the wire quality is deemed non-compliant.

4. The automated serial quality inspection method for heterogeneous flexible wires according to claim 1, characterized in that, The improved gradient function, combined with the width, height, and pixel values ​​of each image in the multi-dimensional image data, calculates the sharpness of the axial edge of the wire, specifically as follows: Determine the target image corresponding to the terminal area and the heat shrink tubing area in the multi-dimensional image data; Select two pixels with a fixed pixel interval along the axial direction in the target image line by line, read the gray values ​​of the two pixels and calculate the absolute value of the difference between the gray values; The absolute values ​​of the grayscale differences across all rows of the target image are summed to obtain the total grayscale difference along the axial edges. Based on the preset effective width of the target image, and combined with the sum of the grayscale differences of the axial edges, the sharpness of the axial edges of the wire is calculated.

5. A serial automated quality inspection device for heterogeneous flexible wires, characterized in that, include: The first acquisition module is used to acquire production order data; The instruction issuing module is used to issue process control instructions to the preset wire processing equipment according to the production order data, and to obtain the heterogeneous soft wire to be tested based on the process control instructions; The second acquisition module is used to control a preset CCD vision system to align with a preset detection area of ​​the heterogeneous flexible wire to be detected, and to acquire multi-dimensional image data of the wire to be detected, specifically: A preset detection area is determined for the heterogeneous flexible wire to be tested, and the preset detection area covers the wire surface, the terminal riveting part and the heat shrink tubing marking part; After controlling the preset CCD vision system to move to a position that matches the preset detection area, the initial imaging sharpness is calculated using a sharpness evaluation method based on the axial edge of the wire. Based on the initial image sharpness, combined with image contrast, the proportion of overexposed areas and the proportion of underexposed areas, a comprehensive quality error is constructed. Based on the comprehensive quality error, the exposure and gain of the CCD vision system are adjusted by a fuzzy PID controller until the comprehensive quality error is less than a preset threshold, thereby obtaining imaging parameters that meet the requirements of the adaptive detection. Based on the imaging parameters and the preset CCD vision system, images are acquired in the preset detection area, and images of the wire surface, terminal riveting part and heat shrink tubing marking part are acquired in sequence to form multi-dimensional image data of the wire to be detected. The feature extraction module is used to extract feature parameters of the wire surface, terminals, and heat shrink tubing from the multi-dimensional image data based on the multi-dimensional image data, combined with an improved gradient function, a preset texture analysis method based on a 3D surface deformation and optical reflection model, and a preset feature matching algorithm, to obtain pixel-level defect detection results for heterogeneous flexible wires. Specifically: Based on the improved gradient function, the sharpness of the axial edge of the wire is calculated by combining the width, height and pixel value of each image in the multi-dimensional image data. Based on the clarity, extract the coordinates of the contour edge points of the terminal area and the boundary line of the heat shrink tubing area to obtain the terminal edge features and the heat shrink tubing boundary features. Based on a preset 3D surface deformation model and combined with a preset adaptive data balancing method, the morphological characteristics of surface scratches are simulated in the wire surface region of multi-dimensional image data. Based on the morphological characteristics and combined with a preset optical reflection model, the reflection intensity of each pixel on the wire surface is calculated. The reflection intensity is compared with the preset reflection intensity of a normal wire surface to identify the texture abnormal area corresponding to the surface scratch, thus forming the texture abnormal feature of the wire surface. Using a preset feature matching algorithm, the terminal edge features are compared with preset terminal standard contour parameters, the heat shrink tubing boundary features are compared with preset heat shrink tubing standard boundary parameters, and the wire surface texture anomaly features are compared with preset wire surface standard reflection parameters. Based on the comparison results, the pixel-level location, defect type, and deviation degree of the defect are determined, and the pixel-level defect detection results of the heterogeneous soft wire are obtained. The determination module is used to decode the pixel-level defect detection results and extract the actual wire number and terminal shape of the wire to be detected; The actual wire number and terminal shape are compared with the production order data to obtain the wire quality compliance judgment result.

6. The automated quality inspection device for heterogeneous flexible wires according to claim 5, characterized in that, The pixel-level defect detection results are decoded to extract the actual wire number and terminal shape of the wire to be inspected; the actual wire number and terminal shape are compared with the production order data to obtain the wire quality compliance judgment result, specifically: Decode the heat shrink tubing identification image in the pixel-level defect detection results to extract the actual wire number of the wire to be inspected; Morphological analysis is performed on the terminal images in the pixel-level defect detection results to extract the terminal morphology of the wire to be inspected; The extracted actual wire number is compared with the preset wire number in the production order data, and the extracted terminal shape is compared with the standard shape corresponding to the terminal model in the production order data. The compliance of the wire quality is determined based on the results of two comparisons. If the actual wire number is consistent with the preset wire number and the terminal shape conforms to the standard shape, the wire quality is deemed compliant; if either comparison result is inconsistent, the wire quality is deemed non-compliant.

Images