Composite conductive fiber quality detection system

Abstract

The invention relates to the field of composite conductive fibers, and discloses a composite conductive fiber quality detection system, which is characterized in that serial number identification and appearance contour extraction are carried out on composite conductive fibers entering a detection area by combining an image identification algorithm with an optical imaging technology; performing electrical property parameter acquisition on the composite conductive fiber by using a non-contact electrical property sensor, and correcting detection data in combination with a temperature compensation algorithm; based on the multispectral imaging and microstructure modeling technology, the integrity and uniformity of the internal and surface structures of the fibers and the adhesion state of a metal layer are analyzed; performing feature fusion analysis on the acquired electrical performance parameters by using a machine learning model, classifying and identifying typical defects, performing grade evaluation on the severity of the typical defects, judging whether the evaluation grade is a high-grade defect, and if so, entering a feedback calibration module; and a targeted regulation and control instruction is automatically generated in combination with a detection result and historical quality data of the production line. The method has the advantage of improving the detection efficiency.

Description

Technical Field

[0001] The present invention relates to the field of composite conductive fibers, and in particular to a composite conductive fiber quality detection system. Background Art

[0002] In the production and application process of composite conductive fibers, ensuring the consistency of their electrical properties and structure is a prerequisite for their reliable application in the fields of smart fabrics, flexible electronic devices and sensors. However, the existing technology still has significant deficiencies in quality inspection. Traditional detection methods mainly rely on manual sampling and then resistance testing, microscopic observation or physical stretching experiments. Such methods are not only time-consuming and labor-intensive, but also often destructive, and cannot achieve real-time, non-destructive testing of the entire batch of conductive fibers. At the same time, because the conductive fibers themselves usually have a multi-material composite structure, fluctuations in their electrical properties may be caused by various factors such as uneven microstructure, interface peeling, and metal layer fracture. Existing detection methods are difficult to accurately locate the cause of the problem, resulting in low detection accuracy, narrow coverage, feedback lag and other problems. Therefore, it is very necessary to design a composite conductive fiber quality detection system that improves detection efficiency. Summary of the Invention

[0003] In view of the shortcomings of the existing technology, the present invention provides a composite conductive fiber quality detection system, which has the advantages of improving detection efficiency and solving the problems in the above background technology. To achieve the above-mentioned purpose of improving detection efficiency, the present invention provides the following technical solution: a composite conductive fiber quality detection system, comprising: Fiber identification module: Through image recognition algorithm combined with optical imaging technology, the composite conductive fiber entering the detection area is numbered and identified and its appearance contour is extracted to determine whether the composite conductive fiber is the target detection object. If so, it enters the electrical acquisition module; Electrical property acquisition module: Uses a non-contact electrical property sensor to collect electrical property parameters of the composite conductive fiber, and combines the temperature compensation algorithm to correct the detection data to determine whether there are any abnormalities in the electrical property parameters. If so, enters the structural analysis module; Structural Analysis Module: Based on multispectral imaging and microstructure modeling technology, it analyzes the integrity, uniformity, and metal layer adhesion of the fiber's internal and surface structures. It also uses abnormal electrical performance parameters to determine whether the structure has defects. If so, it enters the defect identification module. Defect identification module: This module uses a machine learning model to perform feature fusion analysis on the collected electrical performance parameters, classify and identify typical defects, and assess the severity of these defects to determine whether they are high-level defects. If so, the module enters the feedback calibration module. Feedback calibration module: combines test results with historical quality data of the production line to automatically generate targeted control instructions.

[0004] Preferably, the process of numbering and identifying the composite conductive fibers entering the detection area and extracting their appearance contours is as follows: A high-resolution industrial camera deployed above the detection area is used to obtain in-frame image data of the composite conductive fiber. The trained image recognition algorithm model is used to extract the target edge contour of the fiber in the image, and the width, curvature and alignment consistency indicators are calculated through geometric feature fitting methods. Based on the fiber feature point information extracted from the image, a unique identification code is generated in combination with the numbering rules, and the unique identification code is matched with the target detection object list in the database; If the match is successful, the conductive fiber is determined to be the target detection object.

[0005] Preferably, the process of determining whether the composite conductive fiber is a target detection object is as follows: For the composite conductive fiber identified by number, historical production information and specification files are retrieved from the database to extract the target electrical performance parameter range and appearance structure standard value; Compare the deviation between the currently extracted appearance contour parameters and the historical standard parameters, and quantify the deviation score using a normalized method; If the deviation score is lower than the set appearance error tolerance threshold and the number information is within the allowable detection range, the composite conductive fiber is determined to be the target detection object.

[0006] Preferably, the process of correcting the detection data in combination with the temperature compensation algorithm is: The non-contact electrical property sensor collects resistance, capacitance, and conductance, and records the ambient temperature at the detection point and the fiber surface temperature; The thermal response compensation model is used to calculate the theoretical offset of electrical performance indicators under different temperature conditions; Based on the temperature correction coefficient output by the compensation model, the original electrical performance data is numerically corrected to generate a stable and consistent corrected data set.

[0007] Preferably, the process of determining whether the electrical performance parameters are abnormal is as follows: Compare the temperature-compensated electrical property correction data with the historical electrical property standard value of the fiber number and calculate the deviation percentage; Use statistical analysis models to assess whether the current revised data exceeds the historical statistical upper and lower limits, and mark the abnormal factor level; Combining the real-time appearance information of the fiber with the information of the process section, a judgment matrix is constructed to form an abnormality judgment model based on multi-feature weighting; When the abnormal factor level is higher than the set threshold value and the abnormal probability output by the comprehensive judgment model is greater than the confidence threshold, it is determined that the fiber has abnormal electrical properties.

[0008] Preferably, the process of analyzing the integrity and uniformity of the fiber internal and surface structure and the metal layer adhesion state is as follows: A multispectral imaging system is used to perform multi-angle imaging of the composite conductive fiber, obtaining a composite image set including visible light, near-infrared and ultraviolet bands; The arrangement mode, distribution density and metal layer interface parameters of the fiber structure units are extracted through the microstructure image modeling algorithm; Combined with the identified abnormal electrical data, analyze whether there are signs of structural defects; Structural parameters are mapped and analyzed with electrical performance deviation characteristics to form a structural integrity score.

[0009] Preferably, the process of judging whether the structure has defects by combining abnormal electrical performance parameters is as follows: Feature matching is performed on abnormal electrical performance samples and structural modeling results to build an abnormal structure fingerprint database; The multi-factor matching algorithm is used to retrieve the defect type feature vector of the current fiber in the abnormal fingerprint database; Calculate the structural defect probability value based on the vector matching results and compare it with the set identification confidence interval; If the structural defect probability value is greater than the set threshold, it is determined that the abnormal electrical performance fiber structure has defects; If the structural defect probability value is less than or equal to the set threshold, it is determined that the abnormal electrical performance fiber structure has no defects.

[0010] Preferably, the process of performing feature fusion analysis on the collected electrical performance parameters is as follows: Extract the response characteristics of multiple electrical performance parameters in the time dimension, combine them with the defect types identified in the structural analysis results, and perform high-dimensional feature encoding on the electrical performance data; Input the high-dimensional encoding vector into the trained machine learning model to output the predicted value of typical defect types; Feature fusion analysis is performed on the feature differences of different defect categories in the electrical performance parameter space.

[0011] Preferably, the defect types identified by the machine learning model are scored according to a preset severity rating system for the degradation of electrical conductivity, degradation of mechanical properties, and process compatibility issues caused by the defects; The comprehensive evaluation grade is generated by combining the current fiber process stage, quality sensitivity level and predicted defect severity score; When the comprehensive assessment level is greater than or equal to the high-risk warning threshold, it is judged as a high-level defect.

[0012] Preferably, the process of automatically generating targeted control instructions is: Retrieve the corresponding production line process parameter settings and historical defect correction measures based on the defect type, high-risk level, and fiber number information identified in the current inspection task; Use the control rule engine to map the current defect pattern into corresponding production parameter adjustment suggestions; Combine the current production rhythm and equipment operating status to generate targeted control instructions.

[0013] Compared with the prior art, the present invention provides a composite conductive fiber quality detection system with the following beneficial effects: The present invention realizes a closed-loop quality control mechanism from fiber feed identification to defect feedback and control by constructing a full-process detection architecture with coordinated linkage of five modules: identification-collection-analysis-evaluation-control. By fully integrating multi-source heterogeneous technical means such as image recognition, multi-spectral imaging, non-contact electrical detection, microstructure modeling and machine learning, it can accurately extract the appearance profile, electrical performance indicators and microstructure characteristics of composite conductive fibers, and combine temperature compensation algorithms with multi-factor fusion analysis models to effectively improve the stability and accuracy of detection data; by constructing a structure-performance anomaly mapping model and a defect level intelligent assessment system, it can achieve early warning and precise identification of high-risk defects; finally, with the help of a control rule engine, the system can automatically generate control instructions based on historical process parameters and defect repair data, and realize rapid closed-loop adjustment of the production process. The system has significant advantages such as high detection accuracy, fast response speed, strong adaptability and high intelligence level. It is suitable for the stringent requirements for quality consistency and defect controllability in the production process of high-end conductive fibers, and significantly improves the product qualification rate and the intelligence level of the manufacturing process. BRIEF DESCRIPTION OF THE DRAWINGS

[0014] Figure 1 It is a structural schematic diagram of the present invention. DETAILED DESCRIPTION

[0015] The following will clearly and completely describe the technical solutions in the embodiments of the present invention in conjunction with the accompanying drawings. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without making creative efforts are within the scope of protection of the present invention.

[0016] Example 1: Please refer to Figure 1 As shown, a composite conductive fiber quality detection system according to an embodiment of the present invention includes: Fiber identification module: Through image recognition algorithm combined with optical imaging technology, the composite conductive fibers entering the detection area are numbered and their appearance contours are extracted to determine whether the composite conductive fibers are the target detection objects. If so, they enter the electrical acquisition module.

[0017] The process of numbering and identifying the composite conductive fibers entering the detection area and extracting their appearance contours in the fiber identification module is as follows: A high-resolution industrial camera deployed above the detection area is used to obtain in-frame image data of the composite conductive fiber. The trained image recognition algorithm model is used to extract the target edge contour of the fiber in the image, and the width, curvature and alignment consistency indicators are calculated through geometric feature fitting methods. Based on the fiber feature point information extracted from the image, a unique identification code is generated in combination with the numbering rules, and then matched with the target detection object list in the database; If the match is successful, the conductive fiber is determined to be the target detection object.

[0018] A high-resolution industrial camera deployed above the inspection area captures real-time image frames of composite conductive fibers as they pass through the inspection area. Multi-frame averaging and overlaying techniques are used to enhance image clarity and edge contrast, ensuring the complete preservation of fiber detail. A convolutional neural network-based image recognition model is used to segment the input image into target regions, extracting the primary edge features and contour information of the composite conductive fibers within the image. Canny edge detection and multi-scale contour fusion are then used to eliminate background interference and accurately restore the fiber's external structure. Based on the edge contours, a Bezier curve fitting method is used to model the fiber's edge trajectory, further extracting consistency metrics such as fiber width, curvature radius, and overall alignment direction. These geometric parameters are used to assess the tension control and consistency of the fiber during the fabrication process. A unique identifier is generated based on the extracted geometric centerline and its timestamp information. A hash coding mechanism is used to avoid duplicate numbering, and this identifier is matched with the target inspection object number registered in the system database. If the generated number successfully matches an entry in the target inspection list, the system automatically determines the current inspection object as a valid target composite conductive fiber. Otherwise, it is marked as a non-target object and the subsequent electrical testing process is skipped.

[0019] The process of determining whether the composite conductive fiber is a target detection object in the fiber identification module is as follows: For the composite conductive fiber identified by number, historical production information and specification files are retrieved from the database to extract the target electrical performance parameter range and appearance structure standard value; Compare the deviation between the currently extracted appearance contour parameters and the historical standard parameters, and quantify the deviation score using a normalized method; If the deviation score is lower than the set appearance error tolerance threshold and the number information is within the allowable detection range, the composite conductive fiber is determined to be the target detection object.

[0020] According to the unique identification code generated by the number recognition module, the corresponding production batch information and specification files are retrieved from the preset database to extract benchmark information including the target electrical performance parameter range and the corresponding appearance structure standard value; the appearance contour parameters extracted by the current image recognition module are compared with the retrieved standard values one by one, and the deviation degree of each structural indicator is obtained by difference calculation, and further interval normalization processing is performed, such as linear mapping to the [0,1] interval to generate a comprehensive deviation score indicator; a target recognition tolerance threshold is set. If the comprehensive deviation score of the current composite conductive fiber is lower than the threshold and the number recognition result is within the range of the target object number list allowed for detection, the system automatically determines that the fiber is a target detection object; if it is determined to be a target detection object, its status is marked as a compliant object and pushed to the subsequent electrical detection process; if the conditions are not met, it is marked as a non-target object and is eliminated or manually reviewed.

[0021] Electrical property acquisition module: Use non-contact electrical property sensors to collect electrical property parameters of composite conductive fibers, and use temperature compensation algorithms to correct the detection data to determine whether there are any abnormalities in the electrical property parameters. If so, enter the structural analysis module.

[0022] The process of correcting the detection data in the electrical acquisition module by combining the temperature compensation algorithm is as follows: While the non-contact electrical performance sensor collects data such as resistance, capacitance, and conductance, it also records the ambient temperature and fiber surface temperature at the detection point; The thermal response compensation model is used to calculate the theoretical offset of electrical performance indicators under different temperature conditions; Based on the temperature correction coefficient output by the compensation model, the original electrical performance data is numerically corrected to generate a stable and consistent corrected data set.

[0023] While the non-contact electrical property sensor collects electrical parameters such as resistance, capacitance, and conductivity of the composite conductive fiber, the integrated ambient temperature sensor and infrared temperature measurement module are used to record the ambient temperature of the detection area and the instantaneous temperature of the fiber surface, respectively, to ensure that the sampling time points are consistent, thereby forming an electrical-temperature correlation data set; based on the pre-constructed thermal response compensation mathematical model, the synchronously acquired temperature parameters are input, and the theoretical drift value or offset of each electrical property parameter under the current temperature conditions is calculated to obtain an estimate of the impact of temperature on the electrical data; based on the relative difference between the theoretical offset and the reference value under standard conditions, a temperature correction coefficient matrix is generated, and different types of electrical data are differentially corrected through linear or nonlinear mapping functions; the original electrical detection data is normalized and corrected according to the corresponding temperature correction coefficient, and the standardized electrical parameter values after temperature compensation are output to construct a stable and consistent correction data set, which is stored in the detection result database.

[0024] The process of judging whether the electrical performance parameters are abnormal in the electrical acquisition module is as follows: Compare the temperature-compensated electrical property correction data with the historical electrical property standard value of the numbered fiber and calculate the deviation percentage; Use statistical analysis models to assess whether the current revised data exceeds the historical statistical upper and lower limits, and mark the abnormal factor level; Combining the real-time appearance information of the fiber with the information of the process section, a judgment matrix is constructed to form an abnormality judgment model based on multi-feature weighting; When the abnormal factor level is higher than the set threshold value and the abnormal probability output by the comprehensive judgment model is greater than the confidence threshold, it is determined that the fiber has abnormal electrical properties.

[0025] The electrical performance data of the composite conductive fiber after temperature compensation correction are compared item by item with the historical electrical performance standard values associated with the corresponding numbers. For indicators such as resistance, capacitance, and conductance, the relative deviation percentage is calculated using standard formulas to obtain the electrical index deviation vector. Based on the historical electrical performance distribution data of the same type of fiber in the database, a statistical analysis model is constructed to obtain the dynamic upper and lower limit threshold intervals of each electrical index, to determine whether the current deviation vector has crossed the boundary, and accordingly assign an abnormal factor level label, such as mild, moderate, or severe. The electrical abnormal factor level, appearance profile parameters, and production process stage of the current fiber are used as input variables to construct a weighted judgment matrix, and a weighted linear model or decision tree model is used to fuse multi-source features to output an abnormality probability value. The abnormality probability value output by the judgment model is compared with the preset confidence threshold. If the abnormal factor level of the current fiber is higher than the set threshold value and the abnormality probability exceeds the confidence threshold, it is determined that the fiber has an electrical performance abnormality, triggering an alarm mechanism and marking the abnormal data.

[0026] Structural Analysis Module: Based on multispectral imaging and microstructure modeling technology, it analyzes the integrity, uniformity and metal layer adhesion of the fiber's internal and surface structures, and combines abnormal electrical performance parameters to determine whether there are defects in the structure. If so, it enters the defect identification module.

[0027] The process of analyzing the integrity, uniformity of the fiber internal and surface structure and the metal layer adhesion state in the structural analysis module is as follows: A multispectral imaging system is used to perform multi-angle imaging of the composite conductive fiber, obtaining a composite image set including visible light, near-infrared and ultraviolet bands; The arrangement mode, distribution density and metal layer interface integrity parameters of fiber structural units are extracted through microstructure image modeling algorithm; Combined with previously identified abnormal electrical data, analyze whether there are signs of structural defects such as surface peeling, uneven metal plating, structural cracks, or impurity embedding; Structural parameters are mapped and analyzed with electrical performance deviation characteristics to form a structural integrity score.

[0028] By arranging a multispectral imaging system, the composite conductive fiber under inspection is imaged synchronously from multiple angles to obtain multi-channel image data covering visible light, near-infrared and ultraviolet bands, forming a composite image set for microstructural analysis; based on the image set input into the microstructural image modeling algorithm, the cross-sectional texture, longitudinal arrangement state and metal layer attachment boundary of the fiber are multi-scale modeled, and multiple parameter indicators including structural unit arrangement, distribution density, metal coating interface continuity, edge integrity and thickness consistency are extracted; combined with the identified abnormal electrical performance data and its deviation pattern, the multispectral image is further compared to see whether there are typical signs of structural anomalies, including but not limited to surface peeling, uneven metal coating, interface delamination, structural cracks, fiber impurity embedding, bubbles or foreign matter interference, and are identified and marked with the help of image residual analysis and edge discontinuity detection algorithms; a mapping relationship model between structural characteristic parameters and electrical deviation factors is established, and a structural integrity scoring system is formed through feature matching and regression analysis methods to make a comprehensive judgment on the degree of defects of the current inspection object at the structural level.

[0029] The process of judging whether the structure has defects by combining abnormal electrical performance parameters in the structural analysis module is as follows: Feature matching is performed on abnormal electrical performance samples and structural modeling results to build an abnormal structure fingerprint database; The multi-factor matching algorithm is used to retrieve the defect type feature vector that is most similar to the current fiber in the abnormal fingerprint database; Calculate the structural defect probability value based on the vector matching results and compare it with the set identification confidence interval; If the probability of structural defects is greater than the set threshold, it is determined that the fiber structure with abnormal electrical properties has defects; If the structural defect probability value is less than or equal to the set threshold, it is determined that the abnormal electrical performance fiber structure has no defects.

[0030] The corresponding microstructure image modeling parameters and electrical deviation characteristics of composite conductive fiber samples marked as having abnormal electrical properties in historical detection are subjected to correlation analysis, representative structural abnormality feature vectors are extracted, and an abnormal structural fingerprint database is constructed, which contains matching samples of various known defect types and typical electrical response forms. When the currently detected fiber has abnormal electrical properties, the system calls the structural modeling parameters and uses a multi-factor feature matching algorithm to retrieve the most similar defect type feature vector in the abnormal structural fingerprint database, and outputs the matching similarity distribution. Based on the matching results, the similarity scores between multi-dimensional features are integrated to calculate the probability value of the current sample belonging to a specific structural defect type in the database as a quantitative criterion for the structural abnormality risk. The structural defect probability value is compared with the preset recognition confidence interval. When it is higher than the set defect recognition threshold, it is determined that the electrical performance abnormality sample has significant defects at the structural level, and a defect type labeling result is formed.

[0031] Example 2: Figure 1 As shown, a composite conductive fiber quality detection system also includes the following modules: Defect identification module: Uses machine learning models to perform feature fusion analysis on the collected electrical performance parameters, classifies and identifies typical defects, and evaluates the severity of typical defects to determine whether the evaluation level is a high-level defect. If so, enters the feedback calibration module.

[0032] The process of feature fusion analysis of the collected electrical performance parameters in the defect identification module is as follows: Extract the response characteristics of multiple electrical performance parameters in the time dimension, including response speed, stable range, fluctuation amplitude and recovery delay; Combined with the defect types identified in the structural analysis results, high-dimensional feature encoding is performed on the electrical performance data; Input the high-dimensional encoding vector into the trained machine learning model to output the predicted value of typical defect types; Feature fusion analysis is performed on the feature differences of different defect categories in the electrical performance parameter space.

[0033] During the non-contact electrical performance acquisition process, the dynamic response characteristics of multiple electrical performance indicators in the time dimension are extracted, including time series characteristic parameters such as initial response speed, stable interval amplitude, signal fluctuation frequency, maximum fluctuation amplitude, and the time delay required to recover to a stable state. The extracted time series features are jointly encoded with the obtained structural analysis results to form a multimodal high-dimensional feature vector. The encoding process uses principal component analysis or embedded feature selection methods to compress redundant feature dimensions and improve expression efficiency. The encoded high-dimensional feature vector is input into a pre-trained machine learning model, which includes a support vector machine, a convolutional neural network, or an extreme gradient boosting tree, etc., to learn the nonlinear mapping relationship between electrical performance response and structural defects, and output the defect type prediction result and probability score of the current detection fiber. Based on the statistical distribution differences of multiple defect categories in the electrical performance parameter space, the characteristic patterns of various structural defects in the electrical performance response are analyzed to achieve effective differentiation and interpretable fusion of different defect categories in the electrical performance space, thereby optimizing the training strategy and recognition accuracy of the subsequent prediction model.

[0034] The process of determining whether the evaluation level is a high-level defect in the defect identification module is as follows: For defect types identified by the machine learning model, the system scores the potential degradation of electrical conductivity, mechanical properties, and process compatibility issues caused by the defects based on a preset severity rating system. The comprehensive evaluation grade is generated by combining the current fiber process stage, quality sensitivity level and predicted defect severity score; When the comprehensive assessment level is greater than or equal to the high-risk warning threshold, it is judged as a high-level defect.

[0035] The defect type results output by the machine learning model are scored in multiple dimensions based on the preset defect severity level system, including the degradation of conductive performance, mechanical performance degradation, and process compatibility issues that may be caused. The score is based on historical measured data and process tolerance standards, and a standardized mapping algorithm is used to output the defect impact score. The defect impact score is weighted and adjusted based on the production process node where the current composite conductive fiber is located, the quality sensitivity level of the node, and the defect impact score to generate a comprehensive assessment level value. The comprehensive assessment level value is compared with the high-risk warning threshold set by the system. If the level value is greater than or equal to the high-risk threshold, the currently detected fiber is judged to be a high-level defect object.

[0036] Feedback calibration module: combines test results with historical quality data of the production line to automatically generate targeted control instructions.

[0037] The process of automatically generating targeted control instructions in the feedback calibration module is as follows: Retrieve the corresponding production line process parameter settings and historical defect correction measures based on the defect type, high-risk level, and fiber number information identified in the current inspection task; Use the control rule engine to map the current defect pattern into corresponding production parameter adjustment suggestions; Combine the current production rhythm and equipment operating status to generate targeted control instructions.

[0038] Based on the defect type, high-risk level and fiber number information identified in the current inspection task, the production line configuration data associated with the current fiber production batch is retrieved from the process database, including key process parameters and historical records of similar defects; historical control plans and parameter adjustment paths for successful defect repairs are simultaneously queried; the constructed control rule engine is used to perform rule matching analysis on the currently identified defect pattern input, and based on the defect process response model, a corresponding set of production parameter adjustment recommendations is generated, including the recommended adjustment amount, adjustment direction and range of affected equipment; the rule engine is constructed by integrating expert knowledge rules and statistical learning models, and has a bidirectional mapping capability of defects-process responses; the feasibility of the recommended parameter adjustment plan is verified and dynamically optimized in combination with the current production line operation status parameters, and finally a control instruction set including target equipment, adjustment amount, execution priority and feedback path is generated; the instructions can be automatically issued to the corresponding equipment module through the industrial control interface, or prompt the operator to perform semi-automatic intervention.

[0039] It should be noted that, in this document, relational terms such as first and second, etc., are used only to distinguish one entity or operation from another entity or operation, and do not necessarily require or imply any actual relationship or order between these entities or operations. Moreover, the terms "comprises," "comprising," or any other variations thereof are intended to cover non-exclusive inclusion, such that a process, method, article, or apparatus that includes a list of elements includes not only those elements but also other elements not explicitly listed, or elements inherent to such process, method, article, or apparatus.

[0040] While embodiments of the present invention have been shown and described, it will be appreciated by those skilled in the art that various changes, modifications, substitutions, and variations may be made to these embodiments without departing from the principles and spirit of the invention, and that the scope of the invention is defined by the appended claims and their equivalents.

Claims

1. A composite conductive fiber quality detection system, characterized in that: include: Fiber identification module: Through image recognition algorithm combined with optical imaging technology, the composite conductive fiber entering the detection area is numbered and identified and its appearance contour is extracted to determine whether the composite conductive fiber is the target detection object. If so, it enters the electrical acquisition module; Electrical property acquisition module: Uses a non-contact electrical property sensor to collect electrical property parameters of the composite conductive fiber, and combines the temperature compensation algorithm to correct the detection data to determine whether there are any abnormalities in the electrical property parameters. If so, enters the structural analysis module; Structural Analysis Module: Based on multispectral imaging and microstructure modeling technology, it analyzes the integrity, uniformity, and metal layer adhesion of the fiber's internal and surface structures. It also uses abnormal electrical performance parameters to determine whether the structure has defects. If so, it enters the defect identification module. Defect identification module: This module uses a machine learning model to perform feature fusion analysis on the collected electrical performance parameters, classify and identify typical defects, and assess the severity of these defects to determine whether they are high-level defects. If so, the module enters the feedback calibration module. Feedback calibration module: combines test results with historical quality data of the production line to automatically generate targeted control instructions.

2. A composite conductive fiber quality detection system according to claim 1, characterized in that: The process of numbering and identifying the composite conductive fibers entering the detection area and extracting their appearance contours is as follows: A high-resolution industrial camera deployed above the detection area is used to obtain in-frame image data of the composite conductive fiber. The trained image recognition algorithm model is used to extract the target edge contour of the fiber in the image, and the width, curvature and alignment consistency indicators are calculated through geometric feature fitting methods. Based on the fiber feature point information extracted from the image, a unique identification code is generated in combination with the numbering rules, and the unique identification code is matched with the target detection object list in the database; If the match is successful, the conductive fiber is determined to be the target detection object.

3. A composite conductive fiber quality detection system according to claim 2, characterized in that: The process of determining whether the composite conductive fiber is the target detection object is as follows: For the composite conductive fiber identified by number, historical production information and specification files are retrieved from the database to extract the target electrical performance parameter range and appearance structure standard value; Compare the deviation between the currently extracted appearance contour parameters and the historical standard parameters, and quantify the deviation score using a normalized method; If the deviation score is lower than the set appearance error tolerance threshold and the number information is within the allowable detection range, the composite conductive fiber is determined to be the target detection object.

4. A composite conductive fiber quality detection system according to claim 3, characterized in that: The process of correcting the detection data in combination with the temperature compensation algorithm is as follows: The non-contact electrical property sensor collects resistance, capacitance, and conductance, and records the ambient temperature at the detection point and the fiber surface temperature; The thermal response compensation model is used to calculate the theoretical offset of electrical performance indicators under different temperature conditions; Based on the temperature correction coefficient output by the compensation model, the original electrical performance data is numerically corrected to generate a stable and consistent corrected data set.

5. A composite conductive fiber quality detection system according to claim 4, characterized in that: The process of judging whether there is an abnormality in the electrical performance parameters is as follows: Compare the temperature-compensated electrical property correction data with the historical electrical property standard value of the fiber number and calculate the deviation percentage; Use statistical analysis models to assess whether the current revised data exceeds the historical statistical upper and lower limits, and mark the abnormal factor level; Combining the real-time appearance information of the fiber with the information of the process section, a judgment matrix is constructed to form an abnormality judgment model based on multi-feature weighting; When the abnormal factor level is higher than the set threshold value and the abnormal probability output by the comprehensive judgment model is greater than the confidence threshold, it is determined that the fiber has abnormal electrical properties.

6. A composite conductive fiber quality detection system according to claim 5, characterized in that: The process of analyzing the integrity, uniformity of the fiber's internal and surface structure and the metal layer's adhesion is as follows: A multispectral imaging system is used to perform multi-angle imaging of the composite conductive fiber, obtaining a composite image set including visible light, near-infrared and ultraviolet bands; The arrangement mode, distribution density and metal layer interface parameters of the fiber structure units are extracted through the microstructure image modeling algorithm; Combined with the identified abnormal electrical data, analyze whether there are signs of structural defects; Structural parameters are mapped and analyzed with electrical performance deviation characteristics to form a structural integrity score.

7. A composite conductive fiber quality detection system according to claim 6, characterized in that: The process of judging whether the structure has defects by combining abnormal electrical performance parameters is as follows: Feature matching is performed on abnormal electrical performance samples and structural modeling results to build an abnormal structure fingerprint database; The multi-factor matching algorithm is used to retrieve the defect type feature vector of the current fiber in the abnormal fingerprint database; Calculate the structural defect probability value based on the vector matching results and compare it with the set identification confidence interval; If the structural defect probability value is greater than the set threshold, it is determined that the abnormal electrical performance fiber structure has defects; If the structural defect probability value is less than or equal to the set threshold, it is determined that the abnormal electrical performance fiber structure has no defects.

8. A composite conductive fiber quality detection system according to claim 7, characterized in that: The process of feature fusion analysis of the collected electrical performance parameters is as follows: Extract the response characteristics of multiple electrical performance parameters in the time dimension, combine them with the defect types identified in the structural analysis results, and perform high-dimensional feature encoding on the electrical performance data; Input the high-dimensional encoding vector into the trained machine learning model to output the predicted value of typical defect types; Feature fusion analysis is performed on the feature differences of different defect categories in the electrical performance parameter space.

9. A composite conductive fiber quality detection system according to claim 8, characterized in that: The process of determining whether the assessment level is a high-level defect is as follows: For defect types identified by the machine learning model, the system scores the conductivity degradation, mechanical property degradation, and process compatibility issues caused by the defects based on a preset severity rating system. The comprehensive evaluation grade is generated by combining the current fiber process stage, quality sensitivity level and predicted defect severity score; When the comprehensive assessment level is greater than or equal to the high-risk warning threshold, it is judged as a high-level defect.

10. A composite conductive fiber quality detection system according to claim 9, characterized in that: The process of automatically generating targeted control instructions is as follows: Retrieve the corresponding production line process parameter settings and historical defect correction measures based on the defect type, high-risk level, and fiber number information identified in the current inspection task; Use the control rule engine to map the current defect pattern into corresponding production parameter adjustment suggestions; Combine the current production rhythm and equipment operating status to generate targeted control instructions.

Images