An automatic identification system for fabric structure

Abstract

This invention relates to the field of interdisciplinary technology of artificial intelligence and computer vision, and discloses an automatic fabric structure recognition system, including a data acquisition module, a production monitoring module, a production data analysis module, an evaluation module, an adjustment module, an execution module, a feedback module, and an upgrade module. The system achieves precise integration of raw material data and historical process data through the data acquisition module; multi-dimensional data acquisition through the production monitoring module; quantitative calculation of key parameters through the production data analysis module; multi-dimensional performance evaluation through the evaluation module; targeted process optimization through the adjustment module; real-time parameter adjustment through the execution module; closed-loop data flow through the feedback module; and continuous system evolution through the upgrade module. This enables the system to possess intelligent control capabilities throughout the entire process, effectively overcoming the limitations of traditional two-dimensional analysis and achieving high precision, high adaptability, and high dynamic response in fabric structure recognition.

Description

Technical Field

[0001] This invention relates to the field of interdisciplinary technology of artificial intelligence and computer vision, specifically to an automatic identification system for fabric structure. Background Technology

[0002] Current mainstream automatic fabric structure recognition systems primarily rely on two-dimensional projection images for analysis, a method with several limitations. Lacking three-dimensional structural analysis capabilities, the systems cannot accurately capture key three-dimensional information such as fabric thickness and interlayer weaving relationships, leading to significant misjudgments of complex structures—for example, the actual weight of a double-layer plush fabric can differ by up to 30% from an estimate based on single-layer vision. Furthermore, the microscopic fluctuations caused by warp tension during actual weaving (peak and trough amplitudes on the order of 0.5 mm) are not considered; existing algorithms can only perform planar approximations, severely limiting their accuracy in precision simulation applications. The system's dynamic adaptability is also weak: on the one hand, preset parameters struggle to cope with process changes caused by batch variations in raw materials (e.g., different fiber lengths in the same type of cotton due to different origins affect shrinkage rate); experimental data shows that for every unit deviation of the raw material micronaire value from the standard value, the recognition error rate increases by 7%. On the other hand, when dealing with new blended materials (such as graphene-modified fibers), the model update cycle is several months long, with the response speed lagging far behind the pace of market demand changes. These technical bottlenecks indicate that traditional two-dimensional image-based analysis frameworks can no longer meet the needs of high-precision, multi-scenario fabric structure recognition, and there is an urgent need to introduce innovative methods such as three-dimensional reconstruction, dynamic modeling, and adaptive learning to overcome existing limitations. Summary of the Invention

[0003] (a) Technical problems to be solved

[0004] To address the shortcomings of existing technologies, this invention provides an automatic fabric structure recognition system with the advantages of multi-dimensional structure perception, dynamic process adaptation, and intelligent self-evolution. It solves the core problems of traditional two-dimensional image analysis being unable to capture three-dimensional features, fixed thresholds being unable to cope with raw material fluctuations, and the lag in response to new materials.

[0005] (II) Technical Solution

[0006] To achieve the above objectives, the present invention provides the following technical solution: an automatic identification system for fabric structure, characterized in that it includes a data acquisition module, a production monitoring module, a production data analysis module, an evaluation module, an adjustment module, an execution module, a feedback module, and an upgrade module;

[0007] The data acquisition module performs key indicator testing on each batch of raw materials and obtains historical process data corresponding to different types of raw materials.

[0008] The production monitoring module captures three-dimensional data and two-dimensional image data of the fabric by installing online monitoring devices at different locations;

[0009] The production data analysis module outputs latitude and longitude density compensation values ​​based on monitoring and collected data. 3D structural feature parameters Abnormal deviation With cumulative error ;

[0010] The evaluation module is based on the calculation results of the production data analysis module, online monitoring data, and preset evaluation standards;

[0011] The adjustment module constructs a process adjustment experience map based on the evaluation results of the evaluation module and combined with historical correction records, and generates adjustment plans for raw material selection, process parameter settings and production process based on the map.

[0012] The execution module receives real-time adjustments from the adjustment module to the loom parameters, drying temperature, fabric tension, and other process parameters of the production line;

[0013] The feedback module analyzes the execution result data and generates feedback information.

[0014] The upgrade module upgrades and optimizes the system's algorithm model, database, and hardware devices based on long-term feedback data provided by the feedback module.

[0015] Preferably, the data acquisition module includes a raw material index detection unit, a historical data retrieval unit, and a database construction unit.

[0016] Preferably, the raw material index detection unit uses a cotton quality tester, fiber length meter, fiber fineness meter, and other testing equipment to detect the micronaire value, short fiber rate, fiber length, fiber fineness, impurity content, and moisture regain of each batch of raw materials; the historical data retrieval unit obtains the historical process data of the corresponding raw materials through the online variety raw material table, process parameter table, and production result table; the database construction unit integrates the variety raw material data and historical process data to establish a variety raw material material characteristic database.

[0017] Preferably, the production monitoring module is equipped with laser scanning sensors and high-resolution industrial cameras at the loom inlet, loom outlet, dryer outlet, and roll-up machine of the production line. The laser scanning sensors capture three-dimensional data of the fabric, the high-resolution industrial cameras monitor two-dimensional image data of the fabric, the automated sampling equipment performs sampling self-inspection on the beginning and end sections of each roll of fabric, and finally, three-dimensional modeling software is used to construct three-dimensional models of the semi-finished and finished fabric samples.

[0018] Preferably, the production data analysis module includes a latitude and longitude density compensation calculation unit, a three-dimensional structural parameter analysis unit, an anomaly deviation assessment unit, and a cumulative error calculation unit, and calculates the latitude and longitude density compensation value sequentially based on online monitoring data, raw material data, and historical data. 3D structural feature parameters Abnormal deviation With cumulative error .

[0019] Preferably, the latitude and longitude density compensation calculation unit calculates the latitude and longitude density compensation value through the mapping relationship between historical process data and current raw material characteristics. The calculation formula is as follows:

[0020] ;

[0021] In the formula, Indicates the latitude and longitude density compensation value. This indicates the measured micronaire value of the current batch of raw materials. This represents the baseline value for the standard Micron value. This represents the standard deviation of Macron values ​​in the historical database. This represents the measured average fiber length. This indicates the target fiber length setting. This indicates the permissible range of fiber length variation. This represents the workshop relative humidity correction factor. This indicates the weighting factor for the impact of Macron value deviation on latitude and longitude density. This indicates the weighting factor for the impact of fiber length differences on warp and weft density. This represents the weighting factor for the influence of relative humidity in the workshop on latitude and longitude density.

[0022] Preferably, the three-dimensional structural parameter analysis unit calculates the volumetric characteristics of the fabric based on laser-scanned three-dimensional point cloud data. The calculation formula is as follows:

[0023] ;

[0024] In the formula, Indicates the volumetric characteristics of the fabric. , These represent the first three-dimensional point cloud data. The X and Y coordinates of each point , These represent the first three-dimensional point cloud data. +1 point's X and Y coordinates, Indicates the first The fabric thickness value corresponding to each point.

[0025] Preferably, the anomaly deviation assessment unit calculates the anomaly deviation based on the self-inspection data of the first and last sections of each roll of fabric and the standard process parameter values. The calculation formula is as follows:

[0026] ;

[0027] In the formula, Indicates the degree of abnormal deviation. This represents the values ​​of various process parameters in the sampling self-inspection data. This represents the total number of process parameters involved in the calculation. This represents the standard values ​​of various process parameters. This indicates the weight of each process parameter. This represents the average of the sum of standard process parameters.

[0028] Preferably, the cumulative error calculation unit calculates the cumulative error based on continuously monitored data. When the threshold is exceeded, a global recalibration procedure is initiated. The calculation formula is as follows:

[0029] ;

[0030] In the formula, Indicates cumulative error. Indicates the first Error value of the second monitoring, Indicates the first The time interval between the current monitoring and the previous monitoring, where ∑ represents the summation operation.

[0031] Preferably, the evaluation module uses latitude and longitude density compensation values. Based on online monitoring data and preset evaluation standards, the accuracy of latitude and longitude density identification is assessed to determine whether the compensated latitude and longitude density is within the acceptable range. The specific steps are as follows:

[0032] S1.1, When the latitude and longitude density compensation value When ∈[-0.5,0.5], the identification parameters are fine-tuned only through software algorithms without changing the hardware settings;

[0033] S1.2, when latitude and longitude density compensation value When ∈(0.5,2]∪[-2,-0.5), adjust the warp feed parameters of the loom and update the weight factor of the compensation formula simultaneously;

[0034] S1.3, when latitude and longitude density compensation value When the value exceeds the range of (−∞,-2)∪(2,+∞), start the raw material re-inspection procedure → recalibrate the testing equipment → correct the raw material index test value and recalculate the compensation value.

[0035] Based on three-dimensional structural feature parameters Based on online monitoring data and preset evaluation standards, the integrity and consistency of the fabric's three-dimensional structure are assessed, and the three-dimensional structural parameters are analyzed to determine whether they meet design requirements. The specific steps are as follows:

[0036] S2.1, When the three-dimensional structural feature parameters When the structure is within the range of ±3σ of the design value, it is considered to have excellent structural uniformity. When it exceeds this range, potential wrinkle risk areas are marked and associated with heat setting temperature adjustment schemes.

[0037] S2.2 When the volume mutation rate is greater than 15%, the system will automatically trigger a weft density anomaly alarm;

[0038] Based on abnormal deviation Based on online monitoring data and preset evaluation standards, a process stability assessment is conducted to determine whether abnormal deviations are within acceptable thresholds. The judgment process is as follows:

[0039] S3.1, When abnormal deviation When ∈[0,0.1], it is considered a normal fluctuation;

[0040] S3.2, When abnormal deviation When the value is ∈(0.1,0.3], the online re-inspection procedure is initiated and the sampling quantity is increased to confirm the incidental factor;

[0041] S3.3, When abnormal deviation When the value is greater than 0.3, process parameter traceability is triggered → production of this batch is suspended → the influencing factors of raw materials / equipment / environment are traced and process parameters are adjusted;

[0042] Based on cumulative error Based on online monitoring data and preset evaluation standards, the long-term operational accuracy of the system is evaluated to determine whether a global recalibration is needed. The specific steps are as follows:

[0043] S4.1, When the cumulative error If <0.5%, maintain the current observation status;

[0044] S4.2, When the cumulative error When the rate reaches 0.8%, the sampling frequency is doubled.

[0045] S4.3, When the cumulative error When the threshold is ≥1%, the system will forcibly trigger the full production line benchmark calibration process and update the standard reference test piece library;

[0046] The adjustment module adjusts the scheme based on the above evaluation results and the historical correction records to construct a process adjustment experience map.

[0047] Compared with the prior art, the present invention provides an automatic identification system for fabric structure, which has the following advantages:

[0048] 1. This invention overcomes the limitations of two-dimensional analysis by capturing three-dimensional data with a laser scanning sensor, supplements planar detail information with a high-resolution industrial camera, and realizes full-process data visualization through automated sampling and three-dimensional modeling. This system solves the problems of traditional methods being unable to capture three-dimensional structural features, having a single data dimension, and insufficient sample representativeness.

[0049] 2. This invention calculates latitude and longitude density compensation values. This is used to quantify the impact of raw material fluctuations and environmental factors on latitude and longitude density, and to perform real-time dynamic corrections. The calculated latitude and longitude density compensation value... Within the range of [-0.5, 0.5], only the identification parameters are fine-tuned through software algorithms, without changing the hardware settings; when the latitude and longitude density compensation value... When the distance is in the range (0.5,2]⋃[-2,-0.5), adjust the warp feed parameters of the loom and update the weighting factor of the compensation formula simultaneously; when the warp and weft density compensation value... When the value exceeds the range of (−∞,-2)⋃(2,+∞), the raw material re-inspection procedure is initiated, the testing equipment is recalibrated, the raw material index test value is corrected, and the compensation value is recalculated. Finally, the system solves the problem that the fixed threshold cannot cope with the fluctuation of raw material batches.

[0050] 3. This invention uses laser-scanned three-dimensional point cloud data to analyze three-dimensional structural parameters and calculate the volumetric characteristics of the fabric. To accurately reconstruct the spatial topological relationship of the yarn, and to predict the deformation behavior under different tensions, the volume characteristics of the fabric are considered. When the design value is within ±3σ, the structure is judged to have excellent uniformity; if it exceeds this range, potential wrinkle risk areas are marked and associated with heat setting temperature adjustment schemes; when the volume change rate is >15%, the system automatically triggers a weft density abnormality alarm, which finally solves the problem of high error rate in double-layer structure identification and enables reliable identification of complex jacquard fabrics.

[0051] 4. This invention utilizes abnormal deviation. The calculation is used to assess the degree of deviation of a single roll of fabric from the standard process, making process abnormalities quantifiable, traceable, and graded for handling. When the degree of deviation is... When the deviation is within the range of [0, 0.1], it is considered a normal fluctuation; only data is recorded without interfering with production. When the deviation is abnormal... When the value is in the range (0.1, 0.3), initiate the online re-inspection procedure, increase the sampling size, and confirm whether the anomaly is due to an isolated incident; when the anomaly deviation is... When the value exceeds the range of (0.3, +∞), the process parameter traceability is triggered, production of that batch is suspended, the influencing factors of raw materials, equipment and environment are traced and the process parameters are adjusted. In the end, the system solves the problem of delayed detection of production process anomalies and realizes early detection and early handling of quality defects on the production line. Attached Figure Description

[0052] Figure 1 This is a system flowchart of the present invention;

[0053] Figure 2 This is a diagram illustrating the system implementation steps of Embodiment 1 of the present invention;

[0054] Figure 3 This is a diagram illustrating the system implementation steps of Embodiment 2 of the present invention. Detailed Implementation

[0055] 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.

[0056] Please see Figures 1-3 An automatic identification system for fabric structure includes a data acquisition module, a production monitoring module, a production data analysis module, an evaluation module, an adjustment module, an execution module, a feedback module, and an upgrade module.

[0057] The data acquisition module performs key indicator testing on each batch of raw materials and obtains historical process data corresponding to different types of raw materials. Finally, it integrates the raw material data and historical process data to establish a database of material properties of raw materials.

[0058] The production monitoring module captures three-dimensional data and two-dimensional image data of the fabric by installing online monitoring devices at different locations, and uses automated sampling equipment to perform self-inspection of the beginning and end sections of each roll of fabric. Finally, it uses three-dimensional modeling software to construct three-dimensional models of the fabric samples.

[0059] The production data analysis module outputs latitude and longitude density compensation values ​​based on monitoring and collected data. 3D structural feature parameters Abnormal deviation With cumulative error ;

[0060] The evaluation module assesses the accuracy, consistency, and stability of fabric structure identification from multiple dimensions based on the calculation results of the production data analysis module, online monitoring data, and preset evaluation standards, and determines whether the current identification results meet production requirements and quality standards.

[0061] Based on the evaluation results from the assessment module and combined with historical correction records, the adjustment module constructs a process adjustment experience map. Based on this map, it generates adjustment plans for raw material selection, process parameter settings, and production flow to guide subsequent production scheduling and process optimization.

[0062] The execution module receives the adjustment plan generated by the adjustment module and the execution instructions for abnormal handling and recalibration issued by the production data analysis module. It adjusts the loom parameters, drying temperature, fabric tension and other process parameters of the production line in real time, and performs sampling re-inspection and equipment calibration operations at the same time.

[0063] The feedback module collects the execution result data from the execution module, including adjusted process parameters, fabric structure identification results, and production quality data. After analyzing the execution result data, it generates feedback information and feeds it back to the production data analysis module, evaluation module, and adjustment module to provide a basis for subsequent calculations, evaluations, and adjustments.

[0064] Based on long-term feedback data provided by the feedback module, the upgrade module analyzes the system's performance under different raw materials, processes, and production environments, identifies the system's performance shortcomings and potential improvement points, and upgrades and optimizes the system's algorithm model, database, and hardware equipment in conjunction with the development trends of new technologies and new materials in the industry, thereby improving the system's adaptability and recognition accuracy.

[0065] The advantages are: by establishing a data acquisition module to achieve accurate integration of raw material data and historical process data, a production monitoring module to achieve multi-dimensional data acquisition, a production data analysis module to achieve quantitative calculation of key parameters, an evaluation module to achieve multi-dimensional performance evaluation, an adjustment module to achieve targeted process optimization, an execution module to achieve real-time parameter adjustment, a feedback module to achieve closed-loop data flow, and an upgrade module to achieve continuous system evolution, the system has the ability to intelligently manage the entire process, effectively breaking through the limitations of traditional two-dimensional analysis and achieving high precision, high adaptability, and high dynamic response in fabric structure identification.

[0066] The data acquisition module includes a raw material index detection unit, a historical data retrieval unit, and a database construction unit.

[0067] The raw material index testing unit uses cotton quality testers, fiber length meters, fiber fineness meters, and other testing equipment to test the micronaire value, short fiber rate, fiber length, fiber fineness, impurity content, and moisture regain of each batch of raw materials; the historical data retrieval unit obtains the historical process data of the corresponding raw materials through online variety raw material tables, process parameter tables, and production result tables; the database construction unit integrates variety raw material data and historical process data to establish a database of the material characteristics of variety raw materials.

[0068] The advantages are: the raw material index detection unit enables comprehensive quantification of key raw material indicators, the historical data retrieval unit enables accurate traceability of process data, and the database construction unit enables systematic integration of data, giving the system a foundation for correlation analysis of raw materials, processes, and results, and providing data support for dynamic process adaptation.

[0069] The production monitoring module installs laser scanning sensors and high-resolution industrial cameras at the loom entrance, loom exit, dryer exit, and fabric roll machine on the production line. The laser scanning sensors capture three-dimensional data of the fabric, including thickness, interlayer weaving relationship, microscopic undulations caused by warp tension, fiber arrangement direction, and yarn twist distribution. The high-resolution industrial cameras monitor two-dimensional image data of the fabric to supplement the three-dimensional data in terms of planar texture details, color distribution, and local defect identification. Automated sampling equipment performs self-inspection on the beginning and end sections of each roll of fabric. Finally, three-dimensional modeling software is used to construct three-dimensional models of the semi-finished and finished fabric samples, realizing multi-dimensional data acquisition and modeling of the fabric structure from macro to micro.

[0070] The advantages are: by capturing three-dimensional data through laser scanning sensors to overcome the limitations of two-dimensional analysis, high-resolution industrial cameras to supplement planar detail information, and automated sampling and three-dimensional modeling to realize the visualization of data throughout the entire process, the system solves the problems of traditional methods being unable to capture three-dimensional structural features, having a single data dimension, and insufficient sample representativeness.

[0071] The production data analysis module includes a latitude and longitude density compensation calculation unit, a three-dimensional structural parameter analysis unit, an anomaly deviation assessment unit, and a cumulative error calculation unit. Based on online monitoring data, raw material data, and historical data, it sequentially calculates the latitude and longitude density compensation value. 3D structural feature parameters Abnormal deviation With cumulative error .

[0072] The latitude and longitude density compensation calculation unit calculates the latitude and longitude density compensation value by mapping historical process data with current raw material characteristics. It is used to correct latitude and longitude density in real time, and its calculation formula is:

[0073] ;

[0074] In the formula, Indicates the latitude and longitude density compensation value. This indicates the measured micronaire value of the current batch of raw materials. This represents the baseline value for the standard Micron value. This represents the standard deviation of Macron values ​​in the historical database. This represents the measured average fiber length (mm). This indicates the target fiber length setting. This indicates the permissible range of fiber length variation. This indicates the relative humidity correction factor (%) in the workshop. This indicates the weighting factor for the impact of Macron value deviation on latitude and longitude density. This indicates the weighting factor for the impact of fiber length differences on warp and weft density. The weighting factor represents the influence of workshop relative humidity on latitude and longitude density. The above three weighting factors were obtained by fitting historical process optimization records using the least squares method.

[0075] The advantage is that it calculates latitude and longitude density compensation values. This is used to quantify the impact of raw material fluctuations and environmental factors on latitude and longitude density, and to perform real-time dynamic corrections. The calculated latitude and longitude density compensation value... Within the range of [-0.5, 0.5], only the identification parameters are fine-tuned through software algorithms, without changing the hardware settings; when the latitude and longitude density compensation value... When the distance is in the range (0.5,2]⋃[-2,-0.5), adjust the warp feed parameters of the loom and update the weighting factor of the compensation formula simultaneously; when the warp and weft density compensation value... When the value exceeds the range of (−∞,-2)⋃(2,+∞), the raw material re-inspection procedure is initiated, the testing equipment is recalibrated, the raw material index test value is corrected, and the compensation value is recalculated. Finally, the system solves the problem that the fixed threshold cannot cope with the batch fluctuation of raw materials.

[0076] The 3D structural parameter analysis unit performs 3D structural parameter analysis based on laser-scanned 3D point cloud data and calculates the volumetric characteristics of the fabric. This includes volumetric characteristics, interlayer weaving parameters, and thickness values, and its calculation formula is as follows:

[0077] ;

[0078] In the formula, Indicates the volumetric characteristics of the fabric. , These represent the first three-dimensional point cloud data. The X and Y coordinates of each point , These represent the first three-dimensional point cloud data. +1 point's X and Y coordinates, Indicates the first The fabric thickness value corresponding to each point.

[0079] The advantages are: it allows for the analysis of 3D structural parameters using laser-scanned 3D point cloud data, and the calculation of the fabric's volumetric characteristics. To accurately reconstruct the spatial topological relationship of the yarn, and to predict the deformation behavior under different tensions, the volume characteristics of the fabric are considered. When the design value is within ±3σ, the structure is judged to have excellent uniformity; if it exceeds this range, potential wrinkle risk areas are marked and associated with heat setting temperature adjustment schemes; when the volume change rate is >15%, the system automatically triggers a weft density abnormality alarm, which finally solves the problem of high error rate in double-layer structure identification and enables reliable identification of complex jacquard fabrics.

[0080] The anomaly deviation assessment unit calculates the anomaly deviation based on the self-inspection data of the first and last sections of each roll of fabric and the standard process parameter values. This is used to determine whether process parameter tracing is triggered, and its calculation formula is:

[0081] ;

[0082] In the formula, Indicates the degree of abnormal deviation. This represents the values ​​of various process parameters in the sampling self-inspection data. This represents the total number of process parameters involved in the calculation. This represents the standard values ​​of various process parameters. This indicates the weight of each process parameter. This represents the average of the sum of standard process parameters.

[0083] The advantage is: through abnormal deviation. The calculation is used to assess the degree of deviation of a single roll of fabric from the standard process, making process abnormalities quantifiable, traceable, and graded for handling. When the degree of deviation is... When the deviation is within the range of [0, 0.1], it is considered a normal fluctuation; only data is recorded without interfering with production. When the deviation is abnormal... When the value is in the range (0.1, 0.3), initiate the online re-inspection procedure, increase the sampling size, and confirm whether the anomaly is due to an isolated incident; when the anomaly deviation is... When the value exceeds the range of (0.3, +∞), the process parameter traceability is triggered, production of that batch is suspended, the influencing factors of raw materials, equipment and environment are traced and the process parameters are adjusted. In the end, the system solves the problem of delayed detection of production process anomalies and realizes early detection and early handling of quality defects on the production line.

[0084] The cumulative error calculation unit calculates the cumulative error based on continuously monitored data. When the threshold is exceeded, a global recalibration procedure is initiated. The calculation formula is as follows:

[0085] ;

[0086] In the formula, Indicates cumulative error. Indicates the first Error value of the second monitoring, Indicates the first The time interval between the current monitoring and the previous monitoring, where ∑ represents the summation operation.

[0087] The advantage is that it accumulates errors. The calculation makes long-term implicit trends explicit, enabling preventative systematic correction when accumulated errors occur. When the error is <0.5%, maintain the current observation status; when the cumulative error... When the error reaches 0.8%, the sampling frequency is doubled; when the cumulative error... When the threshold is ≥1%, the system forcibly triggers the full production line benchmark calibration process and updates the standard reference test piece library. This ultimately solves the chronic drift problem caused by equipment aging and ensures that the recognition accuracy decay does not exceed 3% of the initial value after 72 hours of continuous operation.

[0088] The evaluation module uses latitude and longitude density compensation values. Based on online monitoring data and preset evaluation standards, the accuracy of latitude and longitude density identification is assessed to determine whether the compensated latitude and longitude density is within the acceptable range; based on three-dimensional structural feature parameters... Based on online monitoring data and preset evaluation standards, the integrity and consistency of the fabric's three-dimensional structure are assessed, and the three-dimensional structural parameters are analyzed to determine whether they meet design requirements; based on the degree of abnormal deviation... Based on online monitoring data and preset evaluation standards, process stability is assessed to determine whether abnormal deviations are within acceptable thresholds; based on cumulative error... Based on online monitoring data and preset evaluation standards, the system's long-term operational accuracy is evaluated to determine whether a global recalibration needs to be initiated.

[0089] Based on the evaluation results from the aforementioned assessment module, and combined with historical corrective action records, the adjustment module constructs a process adjustment experience map to adjust the scheme, specifically as follows:

[0090] (1) When the accuracy of warp and weft density identification is not up to standard, the production data analysis module combines the historical adjustment records of similar raw materials in the experience map, recalculates the weighting factor of micronaire value and fiber length raw material index, corrects the warp and weft density compensation formula, and transmits the recalculated values ​​to the adjustment module through the network to adjust the warp feed parameters of the loom.

[0091] (2) When the integrity and consistency of the three-dimensional structure of the fabric are insufficient, the adjustment module optimizes the installation angle and scanning range of the laser scanning sensor of the production monitoring module based on the experience map, and at the same time adjusts the layer fusion algorithm of the three-dimensional modeling software of the production monitoring module to make targeted adjustments to the weaving process corresponding to the abnormal area of ​​interlayer interlacing parameters.

[0092] (3) When the process stability exceeds the threshold, the adjustment module adjusts the abnormal deviation based on the calculated abnormal deviation. The corresponding process parameter traceability path is used to trace the raw material batch, equipment status and other influencing factors that caused the abnormality, and then the drying machine temperature, cloth winding tension and other related process parameters are adjusted. Finally, the causal relationship of this adjustment is recorded.

[0093] (4) When the long-term operation accuracy assessment of the system shows that global recalibration needs to be initiated, the adjustment module refers to the historical recalibration scheme in the experience map to perform accuracy calibration on the laser scanning sensor, high-resolution industrial camera and detection equipment, update the benchmark values ​​in the standard process parameter library, and retrain the latitude and longitude density compensation calculation model.

[0094] The advantages are: targeted optimization is achieved by accurately matching evaluation results with adjustment plans; decision-making efficiency is improved by inheriting historical adjustment wisdom through experience graphs; the scientific nature of adjustments is ensured by multi-dimensional data linkage analysis; and the system adaptability is continuously improved through a closed-loop adjustment mechanism to guide subsequent production scheduling and process optimization.

[0095] Example 1

[0096] Customized production of high-end fashion fabrics (small batches, multiple varieties): Customers require complex jacquard designs, and the raw materials are imported long-staple cotton and silk blends, with extremely high requirements for color fastness, hand feel and pattern fidelity.

[0097] System implementation steps:

[0098] S1. The data acquisition module quickly detects the micronaire value (measured value = 3.8) and fiber length distribution (mean = 42mm) of the new batch of raw materials, and matches them with the process parameters of similar high-end fabrics in the historical database (such as warp and weft density = 60×40). Because the raw material quality is superior and stable, the warp and weft density compensation value is calculated to be +0.3 (falling in the range of [-0.5, 0.5]). Only the identification parameters are finely adjusted by software algorithm, without the need for hardware intervention.

[0099] S2. The production monitoring module uses laser scanning to capture three-dimensional point cloud data of the fabric and finds that the interlayer interlacing angle deviation is ≤2°. Combined with the texture image collected by the industrial camera, the integrity of the pattern is verified. The automated sampling equipment takes 5 samples from the beginning and end of each roll of fabric to model and confirm that the volume feature fluctuation is within the design value ±3σ, and judges that the structural uniformity is excellent.

[0100] S3. The evaluation module shows that all indicators meet the preset standards (latitude and longitude density pass rate 99%, abnormal deviation <0.1), and it can directly enter the mass production stage;

[0101] S4. The adjustment module calls the "Silk Blend" process library in the experience graph to optimize the drying temperature curve to a gradual temperature increase mode of 120℃ to avoid overheating damage to the silk. The execution module synchronizes parameters to the loom control system in real time to ensure tension fluctuation <±1N;

[0102] S5. The feedback module records that the pass rate of this batch of finished products reaches 99.7%, and stores the data in the knowledge base of the upgrade module for rapid adaptation to subsequent similar orders.

[0103] Implementation results: Zero-trial-error delivery of small-batch customized orders was achieved, and customer satisfaction increased to 98%.

[0104] Example 2

[0105] Large-scale manufacturing of industrial canvas (high load, low cost): daily output exceeds 10,000 meters, raw material is ordinary polyester staple fiber, must withstand outdoor sun exposure and mechanical abrasion, with a focus on tensile strength and weight consistency. System implementation steps:

[0106] S1. The data acquisition module detected that the micronaire value of the raw material suddenly dropped to 2.9 (deviating from the benchmark value of 3.5 to Δ=-0.6), triggering the warp and weft density compensation calculation. Due to the large fluctuation in fiber length (σ_L=4mm), the compensation value rose to -1.8 (falling into the (−2,−0.5) range). The system automatically adjusted the warp feed of the loom by +5% and updated the weight factor α1=0.7 to strengthen the weight of the influence of micronaire value.

[0107] S2. The production monitoring module detected local thickness exceeding the standard at the fabric winding machine (the measured value was 0.15mm thicker than the standard). The 3D modeling showed that the yarn twist in this area was insufficient. The layer marking function was immediately activated to guide subsequent processes to strengthen the compaction treatment of this area.

[0108] S3. The cumulative error of the evaluation module alarm reached 0.9% (close to the 1% threshold), indicating that the equipment has chronic wear. The adjustment module implemented a preventive maintenance plan based on the experience chart: the reeds of looms No. 3 and No. 7 were polished and repaired, and the laser sensor reference surface was calibrated simultaneously.

[0109] S4. The execution module completes global recalibration during off-peak hours at night, and the standard deviation of weight in the production data of the next shift narrows from ±4% to ±1.5%.

[0110] S5, the upgraded module analyzes three months of data trends and suggests refining the polyester raw material grouping strategy into a dual-channel management of "new material / recycled material" to reduce batch-to-batch differences.

[0111] Implementation results: The average monthly defect rate decreased by 40%, and unplanned equipment downtime was reduced by 60%.

[0112] The advantages are as follows: As can be seen from the above embodiments 1 and 2, the system can flexibly adjust the operating mode according to the needs of different production scenarios. In high-end customized scenarios, it can achieve high-quality delivery with accurate data collection and fine-tuning capabilities. In large-scale manufacturing scenarios, it can ensure production stability by relying on dynamic compensation and preventive maintenance. This verifies the universality and efficiency of the automatic fabric structure identification system of the present invention in different production modes of multi-variety, small-batch and large-scale, low-cost production. It fully demonstrates the beneficial effects of the system in improving product quality, reducing production costs and enhancing customer satisfaction.

Claims

1. An automatic identification system for fabric structure, characterized in that, It includes a data acquisition module, a production monitoring module, a production data analysis module, an evaluation module, an adjustment module, an execution module, a feedback module, and an upgrade module; The data acquisition module performs key indicator testing on each batch of raw materials and obtains historical process data corresponding to different types of raw materials. The production monitoring module captures three-dimensional data and two-dimensional image data of the fabric by installing online monitoring devices at different locations; The production data analysis module outputs latitude and longitude density compensation values ​​based on monitoring and collected data. 3D structural feature parameters Abnormal deviation With cumulative error ; The evaluation module is based on the calculation results of the production data analysis module, online monitoring data, and preset evaluation standards; The adjustment module constructs a process adjustment experience map based on the evaluation results of the evaluation module and combined with historical correction records, and generates adjustment plans for raw material selection, process parameter settings and production process based on the map. The execution module receives real-time adjustments from the adjustment module to the loom parameters, drying temperature, fabric tension, and other process parameters of the production line; The feedback module analyzes the execution result data and generates feedback information. The upgrade module upgrades and optimizes the system's algorithm model, database, and hardware devices based on long-term feedback data provided by the feedback module. The production data analysis module includes a latitude and longitude density compensation calculation unit, a three-dimensional structural parameter analysis unit, an anomaly deviation assessment unit, and a cumulative error calculation unit. Based on online monitoring data, raw material data, and historical data, it sequentially calculates the latitude and longitude density compensation value. 3D structural feature parameters Abnormal deviation With cumulative error ; The latitude and longitude density compensation calculation unit calculates the latitude and longitude density compensation value by mapping historical process data with current raw material characteristics. The calculation formula is as follows: ； In the formula, Indicates the latitude and longitude density compensation value. This indicates the measured micronaire value of the current batch of raw materials. This represents the baseline value for the standard Micron value. This represents the standard deviation of Macron values ​​in the historical database. This represents the measured average fiber length. This indicates the target fiber length setting. This indicates the permissible range of fiber length variation. This represents the workshop relative humidity correction factor. This indicates the weighting factor for the impact of Macron value deviation on latitude and longitude density. This indicates the weighting factor for the impact of fiber length differences on warp and weft density. This indicates the weighting factor for the influence of relative humidity in the workshop on latitude and longitude density; The three-dimensional structural parameter analysis unit calculates the volumetric characteristics of the fabric based on laser-scanned three-dimensional point cloud data. The calculation formula is as follows: ； In the formula, Indicates the volumetric characteristics of the fabric. , Let X and Y coordinates be the values ​​of the i-th point in the 3D point cloud data, respectively. , These represent the first three-dimensional point cloud data. +1 point's X and Y coordinates, Indicates the first The fabric thickness value corresponding to each point; The anomaly deviation assessment unit calculates the anomaly deviation based on the self-inspection data of the first and last sections of each roll of fabric and the standard process parameter values. The calculation formula is as follows: ； In the formula, Indicates the degree of abnormal deviation. This represents the values ​​of various process parameters in the sampling self-inspection data. This represents the total number of process parameters involved in the calculation. This represents the standard values ​​of various process parameters. This indicates the weight of each process parameter. This represents the mean of the sum of standard process parameters; The cumulative error calculation unit calculates the cumulative error by continuously monitoring data. When the threshold is exceeded, a global recalibration procedure is initiated. The calculation formula is as follows: ； In the formula, Indicates cumulative error. Indicates the first Error value of the second monitoring, Indicates the first The time interval between the current monitoring and the previous monitoring, where ∑ represents the summation operation.

2. The automatic fabric structure identification system according to claim 1, characterized in that: The data acquisition module includes a raw material index detection unit, a historical data retrieval unit, and a database construction unit.

3. The automatic fabric structure identification system according to claim 2, characterized in that: The raw material index detection unit uses a cotton quality tester, fiber length meter, fiber fineness meter, and other testing equipment to detect the micronaire value, short fiber rate, fiber length, fiber fineness, impurity content, and moisture regain of each batch of raw materials. The historical data retrieval unit obtains the historical process data of the corresponding raw materials through the online variety raw material table, process parameter table, and production result table. The database construction unit integrates the variety raw material data and historical process data to establish a variety raw material material characteristic database.

4. The automatic fabric structure identification system according to claim 1, characterized in that: The production monitoring module is equipped with laser scanning sensors and high-resolution industrial cameras at the loom entrance, loom exit, dryer exit, and roll-up machine of the production line. The laser scanning sensors capture three-dimensional data of the fabric, the high-resolution industrial cameras monitor two-dimensional image data of the fabric, the automated sampling equipment performs sampling inspection on the beginning and end sections of each roll of fabric, and finally, three-dimensional modeling software is used to construct three-dimensional models of the semi-finished and finished fabric samples.

5. The automatic fabric structure identification system according to claim 1, characterized in that: The evaluation module is based on latitude and longitude density compensation values. Based on online monitoring data and preset evaluation standards, the accuracy of latitude and longitude density identification is assessed to determine whether the compensated latitude and longitude density is within the acceptable range. The specific steps are as follows: S1.1, When the latitude and longitude density compensation value When ∈[-0.5,0.5], the identification parameters are fine-tuned only through software algorithms without changing the hardware settings; S1.2, when latitude and longitude density compensation value When ∈(0.5,2]∪[-2,-0.5), adjust the warp feed parameters of the loom and update the weight factor of the compensation formula simultaneously; S1.3, when latitude and longitude density compensation value When the value exceeds the range of (−∞,-2)∪(2,+∞), start the raw material re-inspection procedure → recalibrate the testing equipment → correct the raw material index test value and recalculate the compensation value. Based on three-dimensional structural feature parameters Based on online monitoring data and preset evaluation standards, the integrity and consistency of the fabric's three-dimensional structure are assessed, and the three-dimensional structural parameters are analyzed to determine whether they meet design requirements. The specific steps are as follows: S2.1, When the three-dimensional structural feature parameters When the structure is within the range of ±3σ of the design value, it is considered to have excellent structural uniformity. When it exceeds this range, potential wrinkle risk areas are marked and associated with heat setting temperature adjustment schemes. S2.2 When the volume mutation rate is greater than 15%, the system will automatically trigger a weft density anomaly alarm; Based on abnormal deviation Based on online monitoring data and preset evaluation standards, a process stability assessment is conducted to determine whether abnormal deviations are within acceptable thresholds. The judgment process is as follows: S3.1, When abnormal deviation When ∈[0,0.1], it is considered a normal fluctuation; S3.2, When abnormal deviation When the value is ∈(0.1,0.3], the online re-inspection procedure is initiated and the sampling quantity is increased to confirm the incidental factor; S3.3, When abnormal deviation When the value is greater than 0.3, process parameter traceability is triggered → production of this batch is suspended → the influencing factors of raw materials / equipment / environment are traced and process parameters are adjusted; Based on cumulative error Based on online monitoring data and preset evaluation standards, the long-term operational accuracy of the system is evaluated to determine whether a global recalibration is needed. The specific steps are as follows: S4.1, When the cumulative error If <0.5%, maintain the current observation status; S4.2, When the cumulative error When the rate reaches 0.8%, the sampling frequency is doubled. S4.3, When the cumulative error When the threshold is ≥1%, the system will forcibly trigger the full production line benchmark calibration process and update the standard reference test piece library; The adjustment module adjusts the scheme based on the above evaluation results and the historical correction records to construct a process adjustment experience map.

Images