Textile defect detection method and detection system thereof

Abstract

The invention relates to the field of textile detection. The invention discloses a detection system for textile defect detection, which comprises a light source arranged on one side of a textile and used for outputting a detection light beam; the light path adjusting assembly is located in a light path from the light source to the textile; the image acquisition device is arranged on the other side and is used for acquiring an irradiated image signal; the angle adjusting device adjusts irradiation and acquisition angles of the light source and the image acquisition device; a reflecting component is arranged on one side of the light source and is controlled by a control switch to reflect a detection light beam to the textile; and the image processing unit is connected with the image acquisition device, analyzes the image signal and outputs a defect detection result. According to the detection system, light and shade contrast of flaw edges is enhanced through double-beam interference, and fine defects are accurately captured in combination with angle adjustment and multi-light-source adaptation. The detection method adjusts parameters in stages, automatically matches the optimal detection angle and light source intensity, and improves the adaptability of complex fabric detection.

Description

Technical Field

[0001] The present invention relates to the field of textile detection, and in particular to a method and system for detecting textile defects. Background Art

[0002] In the textile production industry, defect detection is a critical step in ensuring product quality. Traditional inspection methods primarily rely on manual visual inspection or single-light source machine vision technology, which presents significant technical bottlenecks. Manual inspection is subject to subjective judgment and fatigue, making it difficult to consistently and accurately identify subtle defects. Furthermore, efficiency is low, making it unable to meet the demands of high-speed production lines. Although single-light source machine vision is automated, it captures images using only a single illumination angle, making it inadequately adaptable to the complex textures and optical properties of textile surfaces. For example, hidden defects in high-density fabrics can be blurred due to a single illumination angle. Deep defects in plush fabrics can be obscured by insufficient light penetration, while smooth fabrics often suffer from reflections that lead to misidentification. Furthermore, traditional systems lack dynamic adjustment mechanisms. For fabrics with different weave structures, such as warp knitting, weft knitting, and satin, frequent manual adjustments of light source intensity and angle are required. Consequently, detection versatility and stability are limited. In particular, the rate of missed detection of defects such as micro-holes (less than 0.5 mm) and superficial floating points is high, making them inadequate for the high-quality and efficient inspection demands of the modern textile industry.

[0003] With the advancement of textile technology, the widespread use of new fabrics has further highlighted the limitations of traditional detection technology. In existing technologies, a single optical path architecture cannot effectively stimulate the optical characteristic differences of complex defects, and lacks systematic utilization of optical path differences and interference effects, resulting in difficulty in achieving breakthroughs in detection sensitivity. At the same time, high-speed and intelligent production lines require detection systems to have real-time dynamic parameter adjustment capabilities. Traditional methods rely on fixed parameters or manual experience, which has become a technical shortcoming that restricts the automation upgrade of the industry. How to achieve high-sensitivity identification of various textile defects, full-process automated control and multi-scenario adaptive detection has become a technical problem that needs to be solved urgently. Summary of the Invention

[0004] In order to address the deficiencies of the prior art, the present invention discloses a method and a detection system for textile defect detection that integrates multi-angle lighting adjustment, multi-light source adaptation, and intelligent image analysis.

[0005] The present invention discloses a detection system for textile defect detection, which includes a light source, which is located on one side of the textile and is used to output a detection light beam to the textile;

[0006] Light path adjustment component, positioned in the light path from the light source to the textile;

[0007] An image acquisition device, located on the other side of the textile, for acquiring image signals formed after the textile is illuminated by the light source;

[0008] An angle adjustment device, which is used to adjust the illumination angle and acquisition angle of the light source and image acquisition device;

[0009] A reflective component is provided on one side of the light source and is used to reflect the detection light beam of the light source to the textile;

[0010] A control switch is provided on the reflective component and is used to turn the reflective component on and off;

[0011] The image processing unit is connected to the image acquisition device and is used to analyze the image signal and output the defect detection result.

[0012] Furthermore, adjusting the angles of the light source and the image acquisition device includes: adjusting the pitch angle and horizontal angle of the light source and the image acquisition device, the pitch angle being the angle between the irradiation and acquisition direction and the horizontal plane, and the horizontal angle being the angle between the irradiation and acquisition direction and the reference vertical plane;

[0013] During the angle adjustment process, a two-dimensional correspondence between the image signal and the pitch angle and the horizontal angle is obtained.

[0014] Furthermore, the process of adjusting the angle includes:

[0015] Adjusting within a first angle range to obtain a first corresponding relationship between image signals and angle changes;

[0016] Determining a second angle range according to a threshold range of a maximum clarity value in the first corresponding relationship;

[0017] Adjusting within a second angle range to obtain a second corresponding relationship between the image signal and the angle;

[0018] The first angle range is greater than the second angle range;

[0019] The optimal detection angle is determined according to the maximum value in the second corresponding relationship.

[0020] Furthermore, the detection system also includes:

[0021] The multi-light source switching device includes multiple light sources of different wavelengths or intensities, which are arranged circumferentially around the rotation axis.

[0022] Furthermore, the reflective assembly includes a plurality of reflective mirrors with different reflectivities, and the reflective assembly can switch between reflective mirrors with different reflectivities.

[0023] Furthermore, the optical path adjustment component includes an aperture, the aperture range of the aperture is 90%-110% of the diameter of the detection beam of the light source, and is used to limit the irradiation range of the detection beam.

[0024] Furthermore, the image processing unit pre-stores a standard defect feature library, which includes standard textile features and several defect features. By comparing the real-time image signal with the standard defect feature library, the defect type and location of the textile are identified.

[0025] The present invention discloses a method for detecting defects in textiles, which includes S1: acquiring an initial state, selecting different light source types according to the material of the textile, turning on the light source so that the intensity of the light source is a standard detection intensity, directly irradiating the light source onto the textile after adjustment by an optical path adjustment component, turning off a control switch to turn off the reflective component, and obtaining a first image signal of the textile in the initial state by an image processing unit;

[0026] S2: First, the illumination intensity of the light source is adjusted to a first intensity, which is less than the standard detection intensity of the light source. Then, a control switch is turned on to cause the reflective component to reflect the detection beam. The reflected detection beam forms an angle with the detection beam on the textile surface. The angle of the light source and the image acquisition device is adjusted by the angle adjustment device to obtain second image signals at different angles.

[0027] S3: Acquire second image signals at different angles, and during the process of adjusting the angle, obtain a corresponding relationship between the image signal output by the image acquisition device and the change in angle;

[0028] S4: Determine the optimal detection angle according to the maximum image clarity or the maximum defect signal intensity in the corresponding relationship;

[0029] S5: Setting the light source and the image acquisition device at the optimal detection angle, adjusting the illumination intensity to a second intensity, which is greater than the standard detection intensity, and continuously detecting the textile;

[0030] S6: During the continuous detection process, the degree of defects of the textile is determined according to the defect signal output by the image processing unit.

[0031] Furthermore, the detection method further includes: rotating and switching different light sources through a multi-light source switching device to obtain image signals of the textile under different light sources

[0032] Beneficial effects of the present invention:

[0033] The present invention significantly improves the ability to identify subtle defects by combining the dual-beam interference principle with intelligent adjustment technology. The detection system uses the optical interference effect formed by direct light and reflected light at the edge of the defect to make the light and dark contrast of the defect area far exceed that of traditional single-light source detection. Even extremely fine holes or hidden structural defects can be clearly displayed. The detection method uses a phased lighting strategy to first pre-adjust the initial characteristics of the defect with low-intensity light, and then use high-intensity light to enhance the interference effect. The angle adjustment device combines coarse and fine adjustment of the pitch angle and horizontal angle with scanning, automatically matching the optimal detection angle for different fabric texture directions and surface characteristics, and increasing the contrast of the defect signal to a peak state, effectively solving the long-standing problems in the industry such as missed detection of small holes and difficulty in identifying shallow surface defects.

[0034] The detection system and method have built an industrial-grade, efficient detection system, achieving closed-loop control from detection to processing. At the hardware level, the non-contact design avoids fabric damage, the aperture accurately controls the light beam range to reduce stray light interference, and the multi-light source switching device and reflectivity-adjustable reflector cover the optical adaptation requirements of fabrics of different materials; at the software level, the image processing unit automatically completes the real-time image defect type identification and position positioning based on the pre-stored standard defect feature library without manual intervention. The detection method supports the high-speed production line rhythm, and the angle adjustment and light source switching are completed in a short time. It can detect and remove defective fabrics in real time online, significantly reducing subsequent processing losses. The phased detection strategy replaces manual experience with data-driven, dynamically adapts to environmental fluctuations such as light source attenuation and fabric tension changes, and ensures detection stability and robustness.

[0035] The present invention forms a quality control solution that is both universal and economical through technology integration and process innovation. The system hardware's circumferential multi-light source layout, reflector switching and high-precision angle adjustment device, combined with the detection method's staged lighting and dynamic correspondence analysis, can not only meet the non-destructive testing requirements of high-end fabrics such as silk and wool, but is also suitable for rapid testing of conventional fabrics such as denim and industrial fabrics. The automated detection mode greatly reduces manual intervention, shortens equipment debugging time, and lowers the threshold for technology application and labor costs. The overall solution breaks through the bottleneck of traditional machine vision detection. Through the deep integration of optical principle innovation and intelligent control technology, it achieves a comprehensive improvement in the sensitivity, adaptability and efficiency of textile defect detection, providing production companies with a high-precision, cost-effective solution with significant industrial application value and market promotion prospects. BRIEF DESCRIPTION OF THE DRAWINGS

[0036] Figure 1 This is a system flow chart of a detection system for textile defect detection in an embodiment of the present application. DETAILED DESCRIPTION

[0037] In order to enable those skilled in the art to better understand the solutions of the present invention, the technical solutions in the specific implementation manner of the present invention will be clearly and completely described below.

[0038] The present invention discloses a detection system for textile defect detection, which includes a light source, which is located on one side of the textile and is used to output a detection light beam to the textile. An optical path adjustment component is located in the optical path from the light source to the textile. An image acquisition device, which is located on the other side of the textile and is used to acquire image signals formed after the light source irradiates the textile. An angle adjustment device, which is used to adjust the irradiation angle and acquisition angle of the light source and the image acquisition device. A reflection component, which is arranged on one side of the light source, and is used to reflect the detection light beam of the light source to the textile. A control switch, which is arranged on the reflection component, and is used to switch the reflection component. An image processing unit, which is connected to the image acquisition device, and is used to analyze image signals and output defect detection results. A system flow chart of a detection system for textile defect detection is shown as follows: Figure 1 As shown in . Figure 1 This is just one of the system's processes. The specific process is set according to the different fabrics.

[0039] The detection system for textile defect detection of the present invention is constructed through dual detection light paths and angle control. The detection light beam emitted by the light source is calibrated by the light path adjustment component and then directly irradiated onto the textile surface at a set initial angle to form a basic lighting light path. The light path adjustment component can be a lens, an aperture, etc. The reflection component is arranged on the same side of the light source. After being activated by the control switch, it reflects part of the detection light beam emitted by the light source onto the textile, forming an incident light with a specific angle with the direct light beam. The reflection component can be a reflector or a diffuse reflection plate. The angle adjustment device synchronously controls the irradiation angle of the light source and the acquisition angle of the image acquisition device to ensure that the incident angle of the direct light and the reflected light on the textile surface can be dynamically adjusted according to the detection requirements. The typical adjustment range is 15°-75°, and the accuracy is ±0.5°. The angle adjustment device can be a rotating bracket driven by a servo motor.

[0040] When textiles contain defects such as pores or cracks, the optical path difference between direct and reflected light at the edge of the defect satisfies the optical coherence condition and triggers an interference effect. The interference significantly alters the light intensity distribution in the defect area, increasing the brightness in areas of constructive interference and decreasing the brightness in areas of destructive interference. This creates light fringes or spots with a contrast ratio 50%-80% higher than the background. Compared to traditional single-light source detection, the defect's signature is significantly enhanced. An image acquisition device receives the light signal transmitted through the fabric from the other side of the textile and simultaneously captures an image containing the defect's interference signature. The acquisition frame rate is ≥50 frames per second, meeting the real-time detection requirements of high-speed production lines. The image acquisition device can be a high-resolution CCD camera. The image processing unit uses edge detection algorithms and frequency domain analysis to extract characteristic parameters such as the position, width, and contrast of the interference fringes. This is then combined with a deep learning model for defect classification. The frequency domain analysis can be performed using a Fourier transform. The edge detection algorithm can be a Canny operator. The deep learning model can be a convolutional neural network.

[0041] The textile defect detection system of this invention utilizes the principle of dual-beam interference to detect fine pore defects with a diameter of 0.1mm or greater. This significantly improves detection efficiency for fabrics prone to hidden defects, such as warp knitted fabrics and high-density silk, solving the long-standing industry problem of missed small pore detection. For example, when inspecting 200D / 34F polyester filament fabric, the detection rate for 0.15mm pores increased from 60% with traditional methods to 99%. An angle adjustment device allows for dynamic adjustment of incident angles within a range of 0°-90°, effectively highlighting defects of varying orientations and depths. For superficial defects such as hairiness and floating points, a small angle difference of 10°-20° is used to enhance surface texture contrast. For deeper structural defects such as yarn breaks and tissue defects, a large angle difference of 40°-60° is used to amplify the interference path difference of internal defects, increasing the defect signature signal strength by 2-3 times. The entire detection process requires no contact with the textile, eliminating the risk of damage to high-end fabrics associated with contact-based inspection methods such as probes. The integrated design supports online detection and can be seamlessly integrated into the textile production line. The detection speed matches the high-speed weaving process of 200-500m / min, and defective fabrics are removed in real time to reduce subsequent processing losses.

[0042] By combining the dual-beam interference principle with intelligent measurement and control technology, this invention breaks through the technical bottleneck of traditional machine vision inspection in identifying subtle defects, and provides a high-precision and high-efficiency defect detection solution for high-quality textile production.

[0043] In one embodiment, adjusting the angles of the light source and image acquisition device includes adjusting the pitch and horizontal angles of the light source and image acquisition device. The pitch angle is the angle between the illumination and acquisition direction and the horizontal plane, and the horizontal angle is the angle between the illumination and acquisition direction and the reference vertical plane. During the angle adjustment process, a two-dimensional correspondence is obtained, showing how the image signal changes with the pitch and horizontal angles. Pitch adjustment can cover uneven defects at different heights on the textile surface, such as raised points and sunken holes. Horizontal adjustment can capture linear defects with different orientations, such as warp yarn breaks and weft jumps. The multi-angle combination of illumination and acquisition allows each defect to appear more distinctly different in brightness in the image, avoiding the omission of features caused by a single angle. This two-dimensional correspondence directly reflects the intensity changes of the defect signal at different angles. Based on this, the system can automatically match the optimal detection angle combination. For example, for twill fabrics, the horizontal angle can be adjusted to align the grain direction, while for pile fabrics, the pitch angle can be adjusted to reduce surface reflection interference. This eliminates the need for manual trial and error adjustments, improving detection efficiency and stability. Different textiles exhibit significant variations in their light reflection and transmission properties. By recording the correspondence between angles and image signals, the system can dynamically switch detection modes based on fabric characteristics, ensuring clear identification of defect boundaries even in extreme scenarios such as high reflectivity and high transmittance, significantly expanding its scope of application. This two-dimensional correspondence provides multi-dimensional feature input for subsequent image processing. Combined with machine learning algorithms, this allows for the construction of more accurate defect recognition models. For example, by analyzing signal mutation points at specific angles, the interference fringes at the edges of pores can be quickly located, improving the accuracy of identifying subtle defects and providing a basis for automated quality control.

[0044] In one embodiment, the angle adjustment process includes adjusting within a first angle range to obtain a first correspondence between image signal variations and angles. A second angle range is determined based on a threshold range of a maximum clarity value in the first correspondence. Adjusting within the second angle range to obtain a second correspondence between image signal variations and angles. The first angle range is greater than the second angle range. An optimal detection angle is determined based on the maximum value in the second correspondence.

[0045] The present invention's detection system for textile defect detection utilizes a phased adjustment strategy from coarse positioning to fine calibration, combined with a definition threshold to screen key angle ranges. This implementation achieves multiple improvements in detection efficiency, accuracy, and adaptability. Coarse adjustment begins by scanning within a wide first angular range, such as 0°-90° in pitch and 0°-180° in horizontal. By analyzing the overall distribution of image signal clarity, invalid angle regions, such as low-contrast regions with blurred signals, are quickly eliminated. The fine adjustment range is then narrowed to one-third to one-fifth of the original range, reducing the number of angle adjustments required. Fine adjustment then performs high-precision fine adjustments within a focused second angular range, such as the 30°-50° pitch angle range determined by coarse adjustment. This avoids blindly traversing all angles, reduces positioning errors at the optimal detection angle, and significantly improves the definition and contrast of defect images. The optimal detection angles vary significantly for different textiles, such as high-density satin and pile knit fabrics. The coarse adjustment phase identifies the distribution of maximum clarity values, such as the peaks for highly reflective fabrics concentrated at low pitch angles, and automatically defines a fine adjustment range tailored to the specific fabric. For example, when inspecting velvet, the system uses coarse tuning to eliminate interference from strong reflective areas at high pitch angles >60°, directly focusing on the valid angle range without manual intervention. When factors such as light source attenuation and fabric tension fluctuations cause inspection conditions to fluctuate, real-time signal feedback during the coarse tuning phase can redefine the secondary angle range, ensuring stable inspection performance over long periods of operation or during changing operating conditions, eliminating the need for frequent manual calibration. By setting a clarity threshold, such as retaining only angles with signal clarity >60%, the coarse tuning process eliminates ineffective areas such as ambient light interference and fabric texture noise. For example, when inspecting jacquard fabrics, the system automatically avoids angles that could be confused with the pattern texture, allowing the fine tuning phase to focus on true defect features and minimize false positives. The fine tuning phase captures signal abrupt changes within a very small angular range, such as the peak inflection point of the clarity curve. This allows the system to precisely locate the angles that most significantly induce optical interference, maximizing the light-dark contrast at the edges of 0.1mm pores and addressing the problem of missed detections often encountered with single-angle inspection. The phased angle adjustment strategy adopts a hierarchical mechanism of global scanning and positioning followed by local fine calibration, which significantly improves efficiency while ensuring detection accuracy. It has both adaptability to complex scenarios and industrial-grade practicality, providing key technical support for the automation and intelligence of textile defect detection.

[0046] As an embodiment, the detection system also includes a multi-light source switching device, which includes multiple light sources of different wavelengths or intensities, arranged circumferentially around the rotating axis. The multi-light source switching device configured in the detection system can automatically switch to the appropriate light source according to the material, thickness and surface characteristics of the textile by arranging light sources of different wavelengths or intensities circumferentially around the rotating axis. Short-wavelength light sources are suitable for highlighting tiny holes in high-density fabrics, long-wavelength light sources are conducive to penetrating plush fabrics to identify deep defects, and light sources of different intensities can balance the reflective interference of smooth fabrics and the signal attenuation of translucent fabrics. The circumferential layout enables each light source to project light from different angles, and combines with the angle adjustment device to form a multi-dimensional lighting combination, which significantly enhances the optical characteristic differences of various defects. There is no need to manually change the light source or adjust the parameters, and automatic and accurate detection of various fabrics such as knitted fabrics, silk, denim, etc. can be achieved, greatly improving the universality and detection efficiency of the system.

[0047] As an embodiment, the reflective component includes several reflectors with different reflectivities, and the reflective component can switch between reflectors with different reflectivities. The reflective component of the detection system is equipped with several reflectors with different reflectivities and supports switching, which can flexibly adjust the intensity of the reflected light according to the material characteristics and defect types of the textiles. High-reflectivity reflectors can enhance the energy of the reflected light beam and highlight subtle defects under dark or thick fabrics. Low-reflectivity reflectors can reduce strong light interference and clearly present shallow surface defects on the surface of smooth fabrics. By automatically switching the reflector that adapts to the current detection needs, combined with the angle adjustment of the direct light from the light source and the reflected light, it can effectively balance the light reflection and transmission characteristics of different fabrics, forming a differentiated optical interference effect, making the light and dark contrast of the edges of the pores, the reflection difference of different-colored yarns and other defect characteristics easier to identify, without the need to manually replace hardware or repeatedly debug parameters, to achieve efficient and accurate detection of a variety of textiles, significantly improving the adaptability and detection stability of the system.

[0048] As an embodiment, the optical path adjustment component includes an aperture, and the aperture range of the aperture is 90%-110% of the detection beam diameter of the light source, which is used to limit the irradiation range of the detection beam. The optical path adjustment component of the detection system integrates an aperture, and its aperture range is highly matched with the detection beam diameter, which can accurately limit the irradiation range. Make the light beam evenly cover the area to be detected, avoiding edge stray light interference caused by too wide a beam, or detection blind spots caused by too narrow a beam. By controlling the beam boundary, the optical interference effect at the surface defects of the textile can be effectively enhanced, making the light and dark changes at the edges of the pores, the light and shadow differences of structural defects and other features more clearly presented, improving the purity and contrast of the signal obtained by the image acquisition device, and providing high-quality data for the subsequent image processing unit to accurately identify defects. No additional optical calibration steps are required, achieving efficient utilization of the detection beam energy and a stable improvement in detection accuracy.

[0049] As an implementation method, the image processing unit pre-stores a standard defect feature library, which includes standard textile features and several defect features. By comparing the real-time image signal with the standard defect feature library, the type and location of the textile defect are identified. The image processing unit has a built-in standard defect feature library, which can realize automated comparison and analysis of real-time collected images by pre-storing standard textile features and multiple defect features. It can quickly identify the type and precise location of defects on the surface or inside of textiles, avoiding the subjectivity and missed detection problems of manual visual inspection. The feature library covers multi-dimensional feature parameters of common defects, supports universal detection of fabrics with different materials and textures, and combines with dynamic matching of real-time image signals to significantly improve the accuracy and efficiency of defect identification, provide instant quality feedback for the production line, and help realize the intelligence of the entire process from defect detection to grading, without the need for manual real-time monitoring or complex algorithm debugging, reducing detection costs and enhancing system stability.

[0050] The present invention discloses a method for detecting textile defects, comprising the following steps: S1: acquiring an initial state, selecting different light source types based on the material of the textile, turning on the light source so that its intensity reaches a standard detection intensity, and then irradiating the textile directly with the light source after adjustment using an optical path adjustment component. A control switch is then turned off, disabling a reflective component, and an image processing unit acquiring a first image signal of the textile in the initial state. S2: first adjusting the illumination intensity of the light source to a first intensity less than the standard detection intensity of the light source, then turning on the control switch so that the reflective component reflects a detection beam, the reflected detection beam forming an angle with the detection beam on the textile surface, and adjusting the angle of the light source and image acquisition device using an angle adjustment device to acquire second image signals at different angles. S3: acquiring the second image signals at different angles, and during the angle adjustment process, acquiring a corresponding relationship between the image signal output by the image acquisition device and the angle. S4: determining an optimal detection angle based on the maximum image clarity or the maximum defect signal intensity in the corresponding relationship. S5: setting the light source and image acquisition device at the optimal detection angle, adjusting the illumination intensity to a second intensity greater than the standard detection intensity, and continuously inspecting the textile. S6: During the continuous detection process, the degree of defects of the textile is determined according to the defect signal output by the image processing unit.

[0051] The textile defect detection method disclosed in the present invention establishes a dual detection benchmark through S1 initial state acquisition. On the one hand, the light source type is adapted according to the textile material to ensure that the basic lighting matches the optical properties of the fabric, avoiding signal distortion caused by improper light source selection. On the other hand, the reflective component is turned off to perform single-beam acquisition of direct light, obtaining a pure background image without interference, providing an original reference for subsequent dual-beam detection. The S1 step effectively strips away the inherent texture of the fabric and ambient light noise, making the subsequent extraction of defect signals more targeted and solving the problem of missing benchmark data in complex fabric detection.

[0052] In the S2 stage, the light source intensity is reduced and the reflective component is activated, forming a low-energy interference combination of direct and reflected light. Low-intensity illumination reduces overexposure of the fabric surface to strong light, preventing reflective saturation of smooth fabrics or overexposure of the signal of translucent fabrics. The introduction of reflected light creates an initial optical path difference at the edge of the defect, allowing the interference characteristics of subtle defects to initially appear in a low-contrast environment. Combined with dynamic scanning using the angle adjustment device, the basic optical response of defects in different orientations can be fully captured without damaging the fabric, laying the signal foundation for accurately locating the optimal detection angle.

[0053] S3-S4 determines the optimal detection angle through the corresponding relationship between angle and signal, establishing an intelligent parameter adjustment mechanism that integrates global scanning, feature screening, and precise locking. During the angle adjustment process, changes in image clarity and defect signal intensity are recorded in real time, and the system automatically identifies the angle range corresponding to the signal peak. For example, for warp defects, fine-tuning the horizontal angle can enhance the interference fringes at the edges. For deep defects, adjusting the pitch angle can amplify the optical path difference. This process requires no human intervention and automatically filters out invalid angles, transforming detection parameter debugging from manual trial and error to data-driven, significantly improving the adaptation speed and angle positioning accuracy of defect detection on different fabrics.

[0054] In the S5 stage, the light source intensity is increased to the second intensity, forming a stable and efficient detection optical path at the optimal angle. High-intensity direct light and reflected light produce a stronger interference effect in the defect area, further enhancing features such as the light and dark contrast at the edges of fine pores and the difference in reflectivity of different-colored yarns, ensuring that fabric defects can still be clearly imaged even during high-speed movement. Stable lighting conditions, combined with a fixed-angle acquisition device, avoid signal fluctuations caused by vibration and tension changes, providing a continuously stable, high-quality image signal for subsequent S6 continuous detection, ensuring the accuracy and real-time judgment of the degree of defects.

[0055] The detection method for textile defect detection disclosed in the present invention is universal, intelligent and efficient through a closed-loop design of initial benchmark establishment-low light pre-adjustment-angle optimization-strong light detection-real-time judgment. The light source type adaptation, reflective component switch and angle dynamic adjustment cover the detection needs of various fabrics such as knitted fabrics, denim, silk, etc., without the need to replace hardware or reset parameters for different products. The detection parameters are automatically determined based on the correspondence between the image signal and the angle and intensity. Combined with the defect feature library comparison of the image processing unit, the entire process from detection to grading is automated, reducing misjudgments and missed detections caused by manual intervention. The stage adjustment strategy greatly shortens the detection preparation time, and the high-intensity continuous detection mode matches the rhythm of the high-speed production line to achieve instant quality control of detection-feedback-rejection, significantly improving the yield rate and detection efficiency of textile production.

[0056] As an implementation method, the detection method also includes: rotating and switching different light sources through a multi-light source switching device to obtain image signals of the textile under different light sources. In the detection method, by rotating and switching different light sources through a multi-light source switching device and obtaining corresponding image signals, multi-dimensional lighting adaptation can be achieved according to the material differences and defect characteristics of the textiles. Short-wavelength light sources can highlight the tiny holes in high-density fabrics, long-wavelength light sources are conducive to penetrating plush fabrics to identify deep defects, and light sources of different intensities can balance the reflective interference of smooth fabrics and the signal attenuation of translucent fabrics. The circumferentially arranged light sources project light from different angles, combined with the comprehensive analysis of multi-angle image signals, can fully capture the optical characteristic differences of various defects, without the need for manual replacement of light sources or debugging parameters, and automatically adapt to the detection needs of various fabrics such as knitted fabrics, silk, and denim, significantly improving the accuracy of defect recognition and the universality of the system, and providing technical support for the automated detection of complex textiles.

[0057] It should be understood that those skilled in the art can make improvements or changes based on the above description, and all such improvements and changes should fall within the scope of protection of the appended claims of the present invention.

Claims

1. A detection system for textile defect detection, characterized in that: include: A light source is located on one side of the textile and is used to output a detection light beam to the textile; Light path adjustment component, positioned in the light path from the light source to the textile; An image acquisition device, located on the other side of the textile, for acquiring image signals formed after the textile is illuminated by the light source; An angle adjustment device, which is used to adjust the illumination angle and acquisition angle of the light source and image acquisition device; A reflective component is provided on one side of the light source and is used to reflect the detection light beam of the light source to the textile; A control switch is provided on the reflective component and is used to turn the reflective component on and off; The image processing unit is connected to the image acquisition device and is used to analyze the image signal and output the defect detection result.

2. A detection system for textile defect detection according to claim 1, characterized in that: Adjusting the angles of the light source and the image acquisition device includes: adjusting the pitch angle and horizontal angle of the light source and the image acquisition device, the pitch angle being the angle between the irradiation and acquisition direction and the horizontal plane, and the horizontal angle being the angle between the irradiation and acquisition direction and the reference vertical plane; During the angle adjustment process, a two-dimensional correspondence between the image signal and the pitch angle and the horizontal angle is obtained.

3. A detection system for textile defect detection according to claim 2, characterized in that: The process of adjusting the angle includes: Adjusting within a first angle range to obtain a first corresponding relationship between image signals and angle changes; Determining a second angle range according to a threshold range of a maximum clarity value in the first corresponding relationship; Adjusting within a second angle range to obtain a second corresponding relationship between the image signal and the angle; The first angle range is greater than the second angle range; The optimal detection angle is determined according to the maximum value in the second corresponding relationship.

4. A detection system for textile defect detection according to claim 1, characterized in that: The detection system also includes: The multi-light source switching device includes multiple light sources of different wavelengths or intensities, which are arranged circumferentially around the rotation axis.

5. A detection system for textile defect detection according to claim 1, characterized in that: The reflective component includes a plurality of reflective mirrors with different reflectivities, and the reflective component can switch between reflective mirrors with different reflectivities.

6. A detection system for textile defect detection according to claim 1, characterized in that: The optical path adjustment component includes an aperture, the aperture range of which is 90%-110% of the diameter of the detection beam of the light source, and is used to limit the irradiation range of the detection beam.

7. A detection system for textile defect detection according to claim 1, characterized in that: The image processing unit pre-stores a standard defect feature library, which includes standard textile features and several defect features. By comparing the real-time image signal with the standard defect feature library, the defect type and location of the textile are identified.

8. A method for detecting textile defects, using the detection system for textile defects as claimed in any one of claims 1 to 7, characterized in that: include: S1: Initial state acquisition: Select different light source types according to the material of the textile, turn on the light source so that the intensity of the light source is the standard detection intensity, adjust the light path through the light path adjustment component, and then directly illuminate the textile. Turn off the control switch to turn off the reflective component, and the image processing unit obtains a first image signal of the textile in the initial state; S2: First, the illumination intensity of the light source is adjusted to a first intensity, which is less than the standard detection intensity of the light source. Then, a control switch is turned on to cause the reflective component to reflect the detection beam. The reflected detection beam forms an angle with the detection beam on the textile surface. The angle of the light source and the image acquisition device is adjusted by the angle adjustment device to obtain second image signals at different angles. S3: Acquire second image signals at different angles, and during the process of adjusting the angle, obtain a corresponding relationship between the image signal output by the image acquisition device and the change in angle; S4: Determine the optimal detection angle according to the maximum image clarity or the maximum defect signal intensity in the corresponding relationship; S5: Setting the light source and the image acquisition device at the optimal detection angle, adjusting the illumination intensity to a second intensity, which is greater than the standard detection intensity, and continuously detecting the textile; S6: During the continuous detection process, the degree of defects of the textile is determined according to the defect signal output by the image processing unit.

9. The method for detecting textile defects according to claim 8, characterized in that: The detection method further includes: rotating and switching different light sources through a multi-light source switching device to obtain image signals of the textile under different light sources.

Images