A machine vision-based real-time correction system for knitting fabric skew

Abstract

The present application relates to the technical field of knitted fabric processing, and discloses a knitted fabric weft skew real-time correction system based on machine vision, which comprises a feeding tension stabilizing module, a machine vision detection module, a central control module, a partitioned accurate correction module and a discharging re-inspection module arranged in series along the transmission direction of the knitted fabric, and further comprises a power module and a data storage module; the present application has high detection precision and fast response speed, adopts multi-camera cooperative acquisition and adaptive image processing algorithm, and combines with a ring-shaped shadowless light source to realize non-contact and high-precision detection of weft skew parameters, with a weft skew angle detection precision of ≤0.1° and a detection response time of ≤50 ms, effectively solving the problems of low precision and easy interference of the existing detection method, and accurately identifying different types of weft skew phenomena; the present application has strong correction pertinence and high precision; realizes closed-loop correction, improves correction stability; has wide adaptability and strong universality.

Description

Technical Field

[0001] This invention belongs to the field of knitted fabric processing technology, specifically a real-time weft skew correction system for knitted fabrics based on machine vision. Background Technology

[0002] During the weaving, dyeing, finishing, and subsequent processing of knitted fabrics, weft skew is easily caused by equipment precision errors, mechanical transmission deviations, uneven fabric tension, and environmental factors. This means the weft yarns are not perpendicular to the warp yarns and are tilted. Weft skew severely affects subsequent cutting and sewing of knitted fabrics, leading to fabric waste, dimensional deviations in finished products, and reduced product quality and yield. Therefore, weft skew correction is a critical step in the processing of knitted fabrics.

[0003] Existing knitted fabric skew correction systems are mainly divided into two categories: manual correction and automatic mechanical correction. Manual correction relies on the operator's experience, observing the skew with the naked eye and manually adjusting the correction mechanism. This method suffers from low detection accuracy, low correction efficiency, high labor intensity, and large human error, making it unsuitable for large-scale continuous production. Automatic mechanical correction systems mostly use photoelectric sensors to detect skew parameters, but they have drawbacks such as limited detection range, susceptibility to fabric texture and color, and insufficient detection accuracy. Furthermore, most systems use a holistic correction approach, which cannot accurately correct for skew differences in different areas, resulting in poor targeted correction.

[0004] With the development of machine vision technology, some weft skew correction systems have begun to incorporate machine vision detection. However, existing machine vision-based weft skew correction systems still have many shortcomings: First, the image processing algorithm is fixed and cannot be adapted to knitted fabrics with different textures and colors, easily leading to inaccurate weft texture extraction and large deviations in weft skew parameter calculation. Second, there is a response lag between detection and correction, making real-time synchronous correction impossible and resulting in poor correction effects. Third, the correction mechanism is poorly designed, often using a single correction roller structure, which cannot achieve precise zone correction and has poor correction effects for localized or bidirectional weft skew. Fourth, there is a lack of a complete closed-loop verification mechanism, making it impossible to detect the correction effect and optimize the correction parameters in a timely manner, easily leading to over-correction or under-correction.

[0005] Furthermore, in the prior art, such as the invention patent with publication number CN111851042B, a fabric weft straightening or weft straightening and pattern straightening mechanism and method are disclosed. It uses an expansion roller for weft skew correction, but the detection accuracy of this mechanism depends on photoelectric detection sensors or a single industrial camera, and the correction method is overall adjustment, which cannot achieve precise zone correction and has a slow response speed. The invention patent with publication number CN113174738A discloses a fabric texture and pattern correction method and weft straightening machine, which uses multiple sleeve correction rollers for local correction, but this equipment does not combine the high-precision detection advantages of machine vision, and lacks a tension stabilization mechanism and a closed-loop re-inspection mechanism, so there is still room for improvement in correction accuracy and stability. The invention patent with publication number CN103866551A discloses a rapid detection method for fabric weft skew based on machine vision, but this method only involves weft skew detection and does not involve the complete correction system design, so it cannot achieve integrated synchronous operation of detection and correction.

[0006] Therefore, developing a machine vision-based real-time knitted fabric skew correction system with high detection accuracy, fast response speed, strong correction targeting, adaptability to various knitted fabrics, and closed-loop correction function has become the key to solving the shortcomings of existing technologies. Summary of the Invention

[0007] In order to overcome the shortcomings of the prior art, the present invention provides a real-time weft skew correction system for knitted fabrics based on machine vision, which effectively solves the problems mentioned in the background art.

[0008] To achieve the above objectives, the present invention provides the following technical solution: a machine vision-based real-time correction system for weft skew of knitted fabrics, comprising a feeding tension stabilization module, a machine vision detection module, a central control module, a zoned precision correction module, and an output re-inspection module arranged in series along the transmission direction of the knitted fabric, and further comprising a power supply module and a data storage module. The power supply module is electrically connected to the feed tension stabilization module, the machine vision inspection module, the central control module, the zoned precision correction module, the discharge re-inspection module, and the data storage module, respectively, to provide stable power supply to each module; The data storage module is bidirectionally electrically connected to the central control module and is used to store detection data, calibration parameters, operation logs and standard parameter thresholds. The machine vision inspection module, the zoned precision correction module, and the unloading re-inspection module are all unidirectionally electrically connected to the central control module. The machine vision inspection module transmits skew detection data to the central control module, the central control module outputs correction control commands to the zoned precision correction module, and the unloading re-inspection module transmits correction effect detection data to the central control module. The central control module outputs correction control commands based on the detection data from the machine vision detection module, and optimizes the correction parameters based on the re-inspection data from the material discharge re-inspection module, thus forming a closed-loop correction control.

[0009] Preferably, the feeding tension stabilization module includes a feeding bracket, an active transmission roller, a driven pressure roller, a tension sensor, a tension regulating motor, and a tension regulating roller. The active transmission roller and the driven pressure roller are rotatably mounted parallel to each other at the front end of the feeding bracket, with the driven pressure roller located directly above the active transmission roller, forming a knitted fabric feeding channel between them. The tension regulating roller is rotatably mounted at the rear end of the feeding bracket, located behind the active transmission roller and arranged parallel to it. The tension sensor is fixedly mounted at the end of the rotating shaft of the tension regulating roller for real-time acquisition of tension data during the knitted fabric transmission process. The tension regulating motor is driven by the rotating shaft of the tension regulating roller and is used to drive the tension regulating roller to move up and down vertically. Both the tension sensor and the tension regulating motor are electrically connected to the central control module. The tension sensor transmits the acquired tension data to the central control module, and the central control module controls the operation of the tension regulating motor based on the tension data to maintain the tension of the knitted fabric transmission within a preset threshold range.

[0010] Preferably, the machine vision inspection module includes an inspection bracket, multiple sets of industrial cameras, a ring-shaped shadowless light source, an image acquisition card, and an image processing unit. The multiple sets of industrial cameras are evenly spaced and fixedly mounted on the crossbeam of the inspection bracket, with the camera lenses perpendicularly facing the knitted fabric transmission surface. The fields of view of adjacent sets of industrial cameras overlap by 10%-15%, ensuring full coverage of the knitted fabric width. The ring-shaped shadowless light source is fixedly mounted below the lens of each industrial camera, coaxially arranged with the camera lens, to provide uniform, shadowless illumination to the knitted fabric surface. The image acquisition card is electrically connected to both the multiple sets of industrial cameras and the image processing unit, transmitting the knitted fabric image data acquired by the industrial cameras to the image processing unit. The image processing unit incorporates an adaptive image preprocessing algorithm, a weft texture extraction algorithm, and a weft skew parameter calculation algorithm to process the image data, extract weft texture features, calculate the weft skew angle, weft skew offset, and weft skew type, and transmit the processed weft skew detection data to the central control module.

[0011] Preferably, the image processing unit operates by including the following steps: Step 1, Image preprocessing: Adaptive Gaussian filtering algorithm is used to remove image noise, histogram equalization algorithm is used to enhance the contrast between weft and warp yarns, and Otsu method is used to automatically determine the binarization threshold to complete the image binarization process. Step 2, weft texture extraction: Morphological opening operation is used to remove isolated noise points in the image, and the straight line features of the weft texture are extracted by the Hough transform algorithm to determine the trajectory of the weft yarn. Step 3: Calculate the weft skew parameter. Based on the warp direction, calculate the angle between each weft yarn and the warp yarn as the weft skew angle. Statistically analyze the distribution of the weft skew angle within the entire width to determine the maximum weft skew angle, the average weft skew angle, and the weft skew offset. Step 4: Latitude slant type identification. Based on the distribution characteristics of latitude slant angles, identify the latitude slant type as unidirectional latitude slant, bidirectional latitude slant, or local latitude slant, and transmit the latitude slant parameters and latitude slant type as detection data to the central control module.

[0012] Preferably, the central control module includes a main control chip, a signal receiving unit, a signal processing unit, a control command generation unit, and a parameter optimization unit; The main control chip is an embedded industrial controller, model STM32H743VIT6; The signal receiving unit is used to receive the skew detection data transmitted by the machine vision inspection module and the re-inspection data transmitted by the material discharge re-inspection module. The signal processing unit is used to filter, amplify and normalize the received detection data, and compare the processed data with a preset standard parameter threshold to determine whether the latitude exceeds the allowable range. The control command generation unit generates corresponding correction control commands based on the processed latitudinal skew detection data and a preset correction algorithm, thereby controlling the operation of the zone precision correction module. The parameter optimization unit calculates the correction error based on the re-inspection data from the discharge re-inspection module, uses a PID algorithm to adaptively adjust the correction parameters, optimizes the control commands, and stores the optimized correction parameters in the data storage module.

[0013] Preferably, the zoned precision correction module includes a correction bracket, multiple independent correction units, a drive motor assembly, and a position sensor; The multiple sets of independent straightening units are evenly spaced along the width of the knitted fabric. Each set of independent straightening units corresponds to a section of the knitted fabric. Each independent straightening unit includes a straightening roller, an angle adjustment shaft, and a displacement adjustment slider. The straightening roller is rotatably mounted on the displacement adjustment slider, and the axis of the straightening roller forms an adjustable angle with the transmission direction of the knitted fabric. The angle adjustment shaft is driven by the rotating shaft of the straightening roller and is used to adjust the tilt angle of the straightening roller. The displacement adjustment slider is slidably mounted on the guide rail of the straightening bracket and is used to adjust the position of the straightening roller along the width of the knitted fabric. The drive motor assembly includes an angle adjustment motor and a displacement adjustment motor. The angle adjustment motor is driven by the angle adjustment shaft, and the displacement adjustment motor is driven by the displacement adjustment slider. The position sensor is fixedly installed at the end of the straightening roller and is used to collect the tilt angle and position data of the straightening roller in real time. Both the drive motor assembly and the position sensor are electrically connected to the central control module. The central control module controls the operation of the drive motor assembly to adjust the angle and position of the correction rollers of each independent correction unit, thereby achieving precise correction in different zones.

[0014] Preferably, the number of independent correction units is 6-12 groups, and the spacing between two adjacent groups of independent correction units is adapted to the width of the knitted fabric to ensure that there are no correction blind spots within the full width; the surface of the correction roller is covered with an elastic rubber layer with a thickness of 3-5mm and anti-slip texture on the surface to increase the friction with the knitted fabric and prevent the knitted fabric from slipping or being damaged during the correction process.

[0015] Preferably, the material discharge re-inspection module includes a re-inspection bracket, a re-inspection camera, a re-inspection light source, and a re-inspection processing unit; The re-inspection camera is fixedly mounted on the re-inspection bracket, with the lens vertically facing the transmission surface of the knitted fabric, and the acquisition range covers the full width of the knitted fabric. The re-inspection light source is fixedly installed below the re-inspection camera to provide uniform illumination for the re-inspection area; The re-inspection processing unit is electrically connected to the re-inspection camera and is used to process the knitted fabric image acquired by the re-inspection camera, calculate the corrected weft skew angle, determine whether the correction effect meets the preset standard, and transmit the re-inspection result and the corrected weft skew data to the central control module. If the re-inspection result is unqualified, the central control module controls the partition precision correction module to perform correction again and records the number of corrections and related parameters.

[0016] Preferably, the data storage module adopts a combination of SD card and cloud storage. The SD card is used to store the test data, calibration parameters and operation logs of the past 30 days locally, while the cloud storage is used to store all data for a long time, supporting data query, export and anomaly tracing. The power module adopts a switching power supply with an input voltage of 220V AC and an output voltage of 12V DC. It has built-in overload protection, short circuit protection and undervoltage protection circuits to ensure stable system operation.

[0017] Preferably, the system further includes a human-machine interaction module, which is bidirectionally electrically connected to the central control module and includes a touch screen and operation buttons; the touch screen is used to display the system operating status, detection data, calibration parameters, and abnormal alarm information; the operation buttons are used to manually set system parameters, start or stop system operation, and manually intervene in the calibration process; the human-machine interaction module also supports parameter import and export functions to facilitate system debugging and maintenance.

[0018] Compared with the prior art, the beneficial effects of the present invention are: 1. This invention has high detection accuracy and fast response speed. It adopts multi-camera collaborative acquisition and adaptive image processing algorithm, combined with a ring shadowless light source, to achieve non-contact, high-precision detection of latitude parameters. The latitude angle detection accuracy is ≤0.1° and the detection response time is ≤50ms. It effectively solves the problems of low accuracy and susceptibility to interference in existing detection methods, and can accurately identify different types of latitude phenomena. 2. The present invention has strong targeting and high precision. It adopts a zoned precision correction structure. Multiple independent correction units can simultaneously correct the latitudinal slant differences in different regions. It is suitable for various scenarios such as unidirectional latitudinal slant, bidirectional latitudinal slant, and local latitudinal slant. The correction response time is ≤100ms, and the correction accuracy is high, avoiding the problems of over-correction or under-correction caused by overall correction. 3. This invention achieves closed-loop calibration, improves calibration stability, and sets up a material re-inspection module. Combined with the parameter optimization unit of the central control module, it forms a closed-loop calibration mechanism of "detection-calibration-re-inspection-optimization". It can optimize calibration parameters in real time, ensure that the calibration effect meets the standard, and effectively improve the processing quality of knitted fabrics. 4. This invention has wide adaptability and strong versatility. It adopts an adaptive image processing algorithm and an adjustable tension stabilization mechanism, which can adapt to knitted fabrics with different textures, colors and thicknesses. It does not require complex adjustments for different fabrics, reducing the difficulty of operation. 5. This invention has a high degree of automation and reduces costs. It achieves full automation of the feeding tension stabilization, weft skew detection, accurate correction, and discharge re-inspection without manual intervention, reducing the labor intensity of operators, reducing human error, improving production efficiency, and reducing processing costs. 6. The system of this invention has high stability and is easy to maintain. The structure of each module is reasonably designed. The power module has multiple built-in protection circuits. Data storage adopts a combination of local and cloud storage, supports data traceability and parameter optimization, and the human-computer interaction module facilitates system operation and debugging, reducing maintenance costs. 7. This invention organically combines multi-camera collaborative detection, adaptive image processing, precise partition correction, and closed-loop verification, solving the technical pain points of existing systems such as asynchronous detection and correction, poor correction targeting, and weak adaptability. Compared with existing technologies, it has significant creativity and practicality and can be widely applied in the knitted fabric processing industry. Attached Figure Description

[0019] The accompanying drawings are provided to further illustrate the invention and form part of the specification. They are used together with the embodiments of the invention to explain the invention and do not constitute a limitation thereof.

[0020] In the attached diagram: Figure 1 This is a system architecture diagram of the present invention; Figure 2 This is a flowchart of the latitudinal skew detection process of the present invention; Figure 3 This is a block diagram of the real-time correction logic of the present invention; Detailed Implementation

[0021] The technical solutions of the present invention will be clearly and completely described below with reference to the accompanying drawings of the embodiments of the present invention. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. All other embodiments obtained by those skilled in the art based on the embodiments of the present invention without creative effort are within the scope of protection of the present invention.

[0022] Depend on Figures 1-3 This invention relates to a machine vision-based real-time skew correction system for knitted fabrics, comprising a feed tension stabilization module, a machine vision detection module, a central control module, a zoned precision correction module, and an output re-inspection module arranged in series along the fabric transport direction. It also includes a power supply module and a data storage module. The power supply module is electrically connected to each module to provide stable power. The data storage module is bidirectionally electrically connected to the central control module to store relevant data. The machine vision detection module, zoned precision correction module, and output re-inspection module are all unidirectionally electrically connected to the central control module, forming a closed-loop correction control.

[0023] 1. Feed tension stabilization module The feeding tension stabilization module is used to eliminate tension fluctuations during the feeding of knitted fabrics, ensuring smooth fabric transmission and providing stable conditions for subsequent testing and correction. It avoids weft skew detection errors and correction deviations caused by uneven tension. It includes a feeding bracket, an active transmission roller, a driven pressure roller, a tension sensor, a tension regulating motor, and a tension regulating roller. The active transmission roller and driven pressure roller are mounted in parallel rotation at the front end of the feeding bracket, forming a feeding channel for traction and fabric transmission. The tension regulating roller is mounted at the rear end of the feeding bracket, parallel to the active transmission roller. The tension sensor is mounted at the end of the tension regulating roller's shaft, collecting tension data in real time and transmitting it to the central control module. The tension regulating motor is connected to the tension regulating roller, and the central control module controls the tension regulating motor to drive the tension regulating roller up and down based on the tension data, maintaining the fabric tension within a preset threshold range (0.5-2.0N).

[0024] 2. Machine vision inspection module The machine vision inspection module is the core unit of the system, used for high-precision, non-contact, real-time detection of the weft skew parameter of knitted fabrics. It includes a detection bracket, multiple industrial cameras, a ring-shaped shadowless light source, an image acquisition card, and an image processing unit. The multiple industrial cameras are evenly spaced, covering the full width of the fabric, with adjacent cameras overlapping by 10%-15% to avoid blind spots. The ring-shaped shadowless light source is coaxial with the camera lenses, providing uniform, shadow-free illumination to avoid detection errors caused by fabric surface reflections or shadows. The image acquisition card transmits the image data acquired by the cameras to the image processing unit. The image processing unit has a built-in adaptive algorithm to perform image preprocessing, weft texture extraction, weft skew parameter calculation, and weft skew type recognition. The specific process is as follows: Step 1, Image Preprocessing: Adaptive Gaussian filtering algorithm is used to remove image noise (filter kernel size is 3×3), histogram equalization is used to enhance the contrast between weft and warp yarns, and the binarization threshold is automatically determined based on Otsu's method to convert the image into a binary image and highlight the weft texture. Step 2, weft texture extraction: Morphological opening operation (structural element is 5×5 rectangle) is used to remove isolated noise points, and the straight line features of the weft texture are extracted by Hough transform algorithm to determine the trajectory of each weft yarn; Step 3, Calculation of weft skew parameters: Based on the warp direction, calculate the angle between each weft yarn and the warp yarn as the weft skew angle, count the distribution of the weft skew angle within the entire width, and determine the maximum weft skew angle, the average weft skew angle, and the weft skew offset (the distance the weft yarn deviates from the vertical direction of the warp yarn). Step 4, Scoliosis type identification: Based on the scoliosis angle distribution characteristics, identify unidirectional scoliosis (scoliosis direction is consistent across the entire width), bidirectional scoliosis (scoliosis direction is opposite at both ends of the width), or local scoliosis (scoliosis exists in some areas), and transmit the scoliosis parameters and type to the central control module.

[0025] The industrial camera uses a CCD industrial camera with a resolution of 1920×1080 pixels and a frame rate of 30fps to ensure the clarity and real-time performance of image acquisition. The image processing unit uses an FPGA chip, model EP4CE6F17C8, to enable the rapid operation of image processing algorithms, with a detection response time of ≤50ms and a latitude angle detection accuracy of ≤0.1°.

[0026] 3. Central control module The central control module is the core control unit of the system, used to receive signals from various modules, complete data processing, control command generation, and parameter optimization, and realize the coordinated work of various modules in the system. It includes a main control chip, a signal receiving unit, a signal processing unit, a control command generation unit, and a parameter optimization unit. The main control chip adopts the embedded industrial controller STM32H743VIT6, which has fast processing speed and high stability. The signal receiving unit receives the skew detection data from the machine vision inspection module and the re-inspection data from the outgoing material re-inspection module. The signal processing unit filters, amplifies, and normalizes the data, and compares it with the preset standard parameter threshold (the allowable range of skew angle is ≤0.5°) to determine whether the skew exceeds the standard. The control command generation unit generates corresponding correction control commands based on the processed skew data and a preset correction algorithm to control the precise correction module actions of the zone. The parameter optimization unit calculates the correction error based on the re-inspection data and uses a PID algorithm (proportional coefficient Kp=5.0-8.0, integral coefficient Ki=0.1-0.3, derivative coefficient Kd=0.5-1.0) to adaptively adjust the correction parameters, optimize the control commands, and ensure correction accuracy.

[0027] 4. Precise Zonal Calibration Module The zoned precision correction module is used to achieve real-time and precise correction of weft skew in knitted fabrics. It targets the weft skew differences in different areas and is the core correction unit of the system. It includes a correction bracket, multiple independent correction units, a drive motor assembly, and position sensors. Multiple independent correction units (6-12 sets) are evenly arranged along the fabric width, each corresponding to a correction zone, eliminating blind spots. Each independent correction unit includes a correction roller, an angle adjustment shaft, and a displacement adjustment slider. The correction roller is rotatably mounted on the displacement adjustment slider, with a 3-5mm thick elastic rubber layer and anti-slip texture to prevent fabric slippage or damage. The angle adjustment shaft is connected to the correction roller's rotating shaft. The transmission connection adjusts the tilt angle of the straightening roller; the displacement adjustment slider is slidably mounted on the guide rail to adjust the position of the straightening roller along the width direction; the drive motor unit includes an angle adjustment motor and a displacement adjustment motor, which drive the angle adjustment shaft and the displacement adjustment slider to move respectively; the position sensor is installed at the end of the straightening roller to collect the angle and position data of the straightening roller in real time and feed it back to the central control module to form a closed-loop control for correction; the central control module controls the synchronous operation of each independent straightening unit, and adjusts the tilt angle (adjustment range 0°-5°) and position of the straightening roller according to the weft skew angle and type of different zones to achieve precise correction of zones, with a correction response time ≤100ms.

[0028] 5. Material discharge re-inspection module The material output re-inspection module is used to detect the corrected weft skew effect, realizing closed-loop verification of the correction effect and ensuring that the weft skew of the knitted fabric meets the processing standards. It includes a re-inspection bracket, a re-inspection camera, a re-inspection light source, and a re-inspection processing unit. The re-inspection camera is mounted on the re-inspection bracket, covering the full width of the fabric, and acquiring images of the corrected fabric. The re-inspection light source provides uniform illumination. The re-inspection processing unit processes the images, calculates the corrected weft skew angle, and determines whether the correction effect meets the standard. If the re-inspection is qualified, the fabric is output normally. If it is unqualified, the re-inspection processing unit transmits the re-inspection data to the central control module. The central control module controls the zone precision correction module to perform re-correction and records the number of corrections and related parameters until it is qualified.

[0029] 6. Power supply module and data storage module The power supply module uses a switching power supply with an input of 220V AC voltage and an output of 12V DC voltage. It has built-in overload, short circuit and undervoltage protection circuits to ensure stable system operation. The data storage module uses a combination of SD card and cloud storage. The SD card stores the test data, calibration parameters and operation logs of the most recent 30 days locally, while the cloud stores all data long-term, supporting data query, export and anomaly traceability, which is convenient for system debugging, maintenance and production quality traceability.

[0030] 7. Human-Computer Interaction Module The system also includes a human-machine interface module, which is bidirectionally electrically connected to the central control module for system operation, parameter setting, and status monitoring. This module includes a touch screen and operation buttons. The touch screen displays the system's operating status, detection data, calibration parameters, and abnormal alarm information. The operation buttons are used to manually set parameters, start / stop the system, and manually intervene in the calibration process. The system also supports parameter import and export for easy system debugging and maintenance.

[0031] The working process of this invention is as follows: S1. Feeding stage: The knitted fabric is fed through the feeding channel of the feeding tension stabilization module. The active transmission roller pulls the fabric to be transported, the driven pressure roller presses the fabric, the tension sensor collects tension data in real time, and the central control module controls the tension adjustment motor to adjust the tension adjustment roller so that the fabric tension is maintained within the preset range and the fabric is transported smoothly. S2, Inspection Stage: The flattened fabric enters the inspection area of ​​the machine vision inspection module. The ring-shaped shadowless light source provides uniform illumination. Multiple industrial cameras simultaneously acquire fabric images. The image acquisition card transmits the image data to the image processing unit. The image processing unit completes image preprocessing, weft texture extraction, weft skew parameter calculation, and weft skew type recognition through adaptive algorithms, and transmits the inspection data to the central control module. S3, Correction Stage: The central control module receives the weft skew detection data and compares it with the preset standard parameters. If the weft skew exceeds the allowable range, the control command generation unit generates a targeted correction control command and sends it to the zoned precision correction module. Each independent correction unit adjusts the tilt angle and position of the correction roller according to the control command to achieve zoned precision correction. The position sensor collects the angle and position data of the correction roller in real time and feeds it back to the central control module to ensure the correction accuracy. S4. Re-inspection stage: The corrected fabric enters the outgoing re-inspection module. The re-inspection camera captures the fabric image, and the re-inspection processing unit calculates the corrected weft skew angle and judges the correction effect. If it is qualified, the fabric is discharged normally. If it is unqualified, the central control module optimizes the correction parameters based on the re-inspection data, and the control zone precision correction module corrects it again until it is qualified. S5. Data storage and monitoring: The detection data, calibration parameters, operation logs and re-inspection results throughout the process are all stored in the data storage module. The human-machine interaction module displays the system operation status and related data in real time, which is convenient for operators to monitor and manage.

[0032] It should be noted that, in this document, relational terms such as "first" and "second" are used only to distinguish one entity or operation from another, and do not necessarily require or imply any such actual relationship or order between these entities or operations. Furthermore, the terms "comprising," "including," or any other variations thereof are intended to cover non-exclusive inclusion, such that a process, method, article, or apparatus that comprises a list of elements includes not only those elements but also other elements not expressly listed, or elements inherent to such process, method, article, or apparatus.

[0033] Although embodiments of the invention have been shown and described, it will be understood by those skilled in the art that various changes, modifications, substitutions and alterations can be made to these embodiments without departing from the principles and spirit of the invention, the scope of which is defined by the appended claims and their equivalents.

Claims

1. A machine vision-based real-time skew correction system for knitted fabrics, characterized in that, It includes a feeding tension stabilization module, a machine vision inspection module, a central control module, a zoned precision correction module, and an output re-inspection module, arranged in series along the knitted fabric transmission direction, as well as a power supply module and a data storage module. The power supply module is electrically connected to the feed tension stabilization module, the machine vision inspection module, the central control module, the zoned precision correction module, the discharge re-inspection module, and the data storage module, respectively, to provide stable power supply to each module; The data storage module is bidirectionally electrically connected to the central control module and is used to store detection data, calibration parameters, operation logs and standard parameter thresholds. The machine vision inspection module, the zoned precision correction module, and the unloading re-inspection module are all unidirectionally electrically connected to the central control module. The machine vision inspection module transmits skew detection data to the central control module, the central control module outputs correction control commands to the zoned precision correction module, and the unloading re-inspection module transmits correction effect detection data to the central control module. The central control module outputs correction control commands based on the detection data from the machine vision detection module, and optimizes the correction parameters based on the re-inspection data from the material discharge re-inspection module, thus forming a closed-loop correction control.

2. The machine vision-based real-time skew correction system for knitted fabrics according to claim 1, characterized in that: The feeding tension stabilization module includes a feeding bracket, an active transmission roller, a driven pressure roller, a tension sensor, a tension regulating motor, and a tension regulating roller. The active transmission roller and the driven pressure roller are rotatably mounted in parallel at the front end of the feeding bracket, with the driven pressure roller located directly above the active transmission roller, forming a knitted fabric feeding channel between them. The tension regulating roller is rotatably mounted at the rear end of the feeding bracket, located behind the active transmission roller and arranged parallel to it. The tension sensor is fixedly mounted at the end of the rotating shaft of the tension regulating roller, used to collect tension data during the knitted fabric transmission process in real time. The tension regulating motor is connected to the rotating shaft of the tension regulating roller and is used to drive the tension regulating roller to move up and down in the vertical direction. The tension sensor and the tension regulating motor are both electrically connected to the central control module. The tension sensor transmits the collected tension data to the central control module. The central control module controls the operation of the tension regulating motor based on the tension data to maintain the tension of the knitted fabric within a preset threshold range.

3. The machine vision-based real-time skew correction system for knitted fabrics according to claim 1, characterized in that: The machine vision inspection module includes an inspection bracket, multiple sets of industrial cameras, a ring-shaped shadowless light source, an image acquisition card, and an image processing unit. The multiple sets of industrial cameras are evenly spaced and fixedly mounted on the crossbeam of the inspection bracket, with the camera lenses perpendicularly facing the knitted fabric transmission surface. The fields of view of adjacent sets of industrial cameras overlap by 10%-15%, ensuring full coverage of the knitted fabric width. The ring-shaped shadowless light source is fixedly mounted below the lens of each industrial camera, coaxially arranged with the camera lens, to provide uniform, shadowless illumination to the knitted fabric surface. The image acquisition card is electrically connected to both the multiple sets of industrial cameras and the image processing unit, transmitting the knitted fabric image data acquired by the industrial cameras to the image processing unit. The image processing unit incorporates an adaptive image preprocessing algorithm, a weft texture extraction algorithm, and a weft skew parameter calculation algorithm to process the image data, extract weft texture features, calculate the weft skew angle, weft skew offset, and weft skew type, and transmit the processed weft skew detection data to the central control module.

4. The machine vision-based real-time skew correction system for knitted fabrics according to claim 3, characterized in that: The image processing unit operates by the following steps: Step 1, Image preprocessing: Adaptive Gaussian filtering algorithm is used to remove image noise, histogram equalization algorithm is used to enhance the contrast between weft and warp yarns, and Otsu method is used to automatically determine the binarization threshold to complete the image binarization process. Step 2, weft texture extraction: Morphological opening operation is used to remove isolated noise points in the image, and the straight line features of the weft texture are extracted by the Hough transform algorithm to determine the trajectory of the weft yarn. Step 3: Calculate the weft skew parameter. Based on the warp direction, calculate the angle between each weft yarn and the warp yarn as the weft skew angle. Statistically analyze the distribution of the weft skew angle within the entire width to determine the maximum weft skew angle, the average weft skew angle, and the weft skew offset. Step 4: Latitude slant type identification. Based on the distribution characteristics of latitude slant angles, identify the latitude slant type as unidirectional latitude slant, bidirectional latitude slant, or local latitude slant, and transmit the latitude slant parameters and latitude slant type as detection data to the central control module.

5. The machine vision-based real-time skew correction system for knitted fabrics according to claim 1, characterized in that: The central control module includes a main control chip, a signal receiving unit, a signal processing unit, a control command generation unit, and a parameter optimization unit. The main control chip is an embedded industrial controller, model STM32H743VIT6; The signal receiving unit is used to receive the skew detection data transmitted by the machine vision inspection module and the re-inspection data transmitted by the material discharge re-inspection module. The signal processing unit is used to filter, amplify and normalize the received detection data, and compare the processed data with a preset standard parameter threshold to determine whether the latitude exceeds the allowable range. The control command generation unit generates corresponding correction control commands based on the processed latitudinal skew detection data and a preset correction algorithm, thereby controlling the operation of the zone precision correction module. The parameter optimization unit calculates the correction error based on the re-inspection data from the discharge re-inspection module, uses a PID algorithm to adaptively adjust the correction parameters, optimizes the control commands, and stores the optimized correction parameters in the data storage module.

6. The machine vision-based real-time skew correction system for knitted fabrics according to claim 1, characterized in that: The precise correction module for the zonal area includes a correction bracket, multiple independent correction units, a drive motor assembly, and a position sensor. The multiple sets of independent straightening units are evenly spaced along the width of the knitted fabric. Each set of independent straightening units corresponds to a section of the knitted fabric. Each independent straightening unit includes a straightening roller, an angle adjustment shaft, and a displacement adjustment slider. The straightening roller is rotatably mounted on the displacement adjustment slider, and the axis of the straightening roller forms an adjustable angle with the transmission direction of the knitted fabric. The angle adjustment shaft is driven by the rotating shaft of the straightening roller and is used to adjust the tilt angle of the straightening roller. The displacement adjustment slider is slidably mounted on the guide rail of the straightening bracket and is used to adjust the position of the straightening roller along the width of the knitted fabric. The drive motor assembly includes an angle adjustment motor and a displacement adjustment motor. The angle adjustment motor is driven by the angle adjustment shaft, and the displacement adjustment motor is driven by the displacement adjustment slider. The position sensor is fixedly installed at the end of the straightening roller and is used to collect the tilt angle and position data of the straightening roller in real time. Both the drive motor assembly and the position sensor are electrically connected to the central control module. The central control module controls the operation of the drive motor assembly to adjust the angle and position of the correction rollers of each independent correction unit, thereby achieving precise correction in different zones.

7. The machine vision-based real-time skew correction system for knitted fabrics according to claim 6, characterized in that: The number of independent correction units is 6-12 sets. The spacing between two adjacent sets of independent correction units is adapted to the width of the knitted fabric to ensure that there are no correction blind spots within the full width. The surface of the correction roller is covered with an elastic rubber layer with a thickness of 3-5mm and anti-slip texture to increase the friction with the knitted fabric and prevent the knitted fabric from slipping or being damaged during the correction process.

8. The machine vision-based real-time skew correction system for knitted fabrics according to claim 1, characterized in that: The material discharge re-inspection module includes a re-inspection bracket, a re-inspection camera, a re-inspection light source, and a re-inspection processing unit; The re-inspection camera is fixedly mounted on the re-inspection bracket, with the lens vertically facing the transmission surface of the knitted fabric, and the acquisition range covers the full width of the knitted fabric. The re-inspection light source is fixedly installed below the re-inspection camera to provide uniform illumination for the re-inspection area; The re-inspection processing unit is electrically connected to the re-inspection camera and is used to process the knitted fabric image acquired by the re-inspection camera, calculate the corrected weft skew angle, determine whether the correction effect meets the preset standard, and transmit the re-inspection result and the corrected weft skew data to the central control module. If the re-inspection result is unqualified, the central control module controls the partition precision correction module to perform correction again and records the number of corrections and related parameters.

9. A machine vision-based real-time skew correction system for knitted fabrics according to claim 1, characterized in that: The data storage module uses a combination of SD card and cloud storage. The SD card is used to store the test data, calibration parameters and operation logs of the past 30 days locally, while the cloud storage is used to store all data long-term, supporting data query, export and anomaly tracing. The power module uses a switching power supply with an input voltage of 220V AC and an output voltage of 12V DC. It has built-in overload protection, short circuit protection and undervoltage protection circuits to ensure stable system operation.

10. The machine vision-based real-time skew correction system for knitted fabrics according to any one of claims 1-9, characterized in that, The system also includes a human-machine interface module, which is bidirectionally electrically connected to the central control module. The human-machine interface module includes a touch screen and operation buttons. The touch screen is used to display the system operating status, detection data, calibration parameters, and abnormal alarm information. The operation buttons are used to manually set system parameters, start or stop system operation, and manually intervene in the calibration process. The human-machine interface module also supports parameter import and export functions to facilitate system debugging and maintenance.

Images