A high-precision detection method for key points of textiles based on ultraviolet light

Abstract

The present invention discloses a high-precision UV-based detection method for textile key points, relating to the field of textile material technology. The method comprises the following steps: determining the UV wavelength, constructing a UV fluorescence spectral feature library, adjusting the UV light source, capturing images of the textile surface, preprocessing the captured UV fluorescence images, extracting key point features, identifying and classifying defects, accurately locating defective areas through image segmentation, and outputting detection results. This UV-based high-precision detection method for textile key points achieves deep fusion of multidimensional features and precise key point location by introducing a spectral texture joint distribution model and a saliency segmentation model. Supported by uniform UV light source illumination and combined with multidimensional saliency analysis, it significantly improves detection robustness and segmentation accuracy in complex backgrounds.

Description

Technical Field

[0001] The present invention relates to the technical field of textile materials, and in particular to a high-precision detection method for key points of textiles based on ultraviolet rays. Background Art

[0002] In textile manufacturing and quality control, high-precision inspection of key points is an important means of identifying surface defects, cracks, and other anomalies in fabrics. Traditional inspection methods rely on visible light image analysis or manual visual inspection, which makes it difficult to accurately identify minor defects and internal structural changes. In recent years, ultraviolet-based inspection technology has become a research hotspot due to its ability to reveal hidden surface features and spectral anomalies in fabrics. However, existing technologies have the following major shortcomings:

[0003] Single feature extraction: Most methods perform detection based on only one dimension of texture, spectrum or edge features, ignoring the complementarity between these features, resulting in insufficient robustness of detection in complex texture backgrounds.

[0004] Low segmentation accuracy: Traditional segmentation methods, such as algorithms based on fixed thresholds or simple edge detection, are difficult to adapt to the complex changes in fabric surface texture and spectral response, resulting in inaccurate defect positioning. Summary of the Invention

[0005] In view of the shortcomings of the existing technology, the present invention provides a high-precision detection method for key points of textiles based on ultraviolet light to solve the problems raised in the above background technology.

[0006] To achieve the above object, the present invention provides the following technical solutions:

[0007] In a first aspect, an embodiment of the present invention provides a high-precision detection method for key points of a textile based on ultraviolet light, comprising the following steps:

[0008] S1. Determine the wavelength of ultraviolet light for detection based on fluorescence spectroscopy analysis;

[0009] S2. Constructing a UV fluorescence spectrum feature library for textiles;

[0010] S3. Conduct a comprehensive analysis of the detection area and adjust the UV light source;

[0011] S4, using a high-resolution UV camera to capture images of the textile surface;

[0012] S5, preprocessing the collected ultraviolet fluorescence image;

[0013] S6, extracting key point features from the preprocessed image;

[0014] S7, identifying and classifying defects using the extracted key point features;

[0015] S8. Based on defect recognition, the defect area is accurately located through image segmentation;

[0016] S9. Output the test results and generate a test report.

[0017] To further optimize the technical solution, in step S1, the surface of the textile is irradiated by an ultraviolet light source to stimulate a fluorescent reaction on the surface of the textile, and a spectral analysis is performed to evaluate the fluorescent response characteristics of the textile after ultraviolet irradiation, and to determine the ultraviolet wavelength for detection.

[0018] To further optimize the technical solution, in step S2, ultraviolet irradiation experiments are performed on textile samples of different batches, materials, and dyeing methods, and their fluorescence spectrum characteristics are recorded to form an ultraviolet fluorescence spectrum characteristic library;

[0019] The feature library contains all variables that may affect the detection accuracy, including factors such as textile material, dye type, and fabric structure.

[0020] To further optimize the technical solution, in step S3, an ultraviolet intensity distribution measuring instrument is used to detect the ultraviolet intensity distribution of the detection area;

[0021] By adjusting the position, angle and irradiation distance of the UV light source, uniform UV distribution can be achieved, ensuring that each inspection point on the textile is evenly irradiated by UV rays.

[0022] To further optimize this technical solution, in step S5, the preprocessing includes the following process:

[0023] First, an image filtering algorithm is applied to remove background noise and eliminate ambient light interference;

[0024] Then, the image enhancement algorithm is used to enhance the contrast of the image and enhance the image of the areas with weak fluorescence, so that the tiny defects on the surface of the textile are clearly presented.

[0025] To further optimize the technical solution, in step S6, when extracting key point features, a spectral texture joint distribution model is constructed to integrate the fluorescence response, texture change and shape characteristics of the textile surface, and coupled using weight factors;

[0026] In the spectral texture joint distribution model, let any pixel point on the textile surface be , its eigenvalue is calculated by the following formula:

[0027] ;

[0028] in,

[0029] : Pixels The comprehensive eigenvalue of is used to indicate the possibility of the point being a key point;

[0030] : The eigenvalue based on texture gradient is used to represent the texture complexity of the pixel;

[0031] : Spectral intensity characteristic value, calculated by the spectral response of the pixel point in the ultraviolet fluorescence image;

[0032] : Shape eigenvalue, calculated by the significance of the pixel in the surrounding geometric structure;

[0033] : Weight factor, indicating the importance of texture, spectrum, and shape features in key point extraction.

[0034] To further optimize this technical solution, in the spectral texture joint distribution model, the calculation of the eigenvalue includes:

[0035] Texture gradient feature calculation:

[0036] Spectral intensity feature calculation;

[0037] Shape feature calculation;

[0038] Weight optimization and feature fusion.

[0039] To further optimize the present technical solution, in step S7, the extracted key point features are input into a machine learning model. Through the training set sample data, the machine learning model learns the characteristics of different types of textile defects and automatically identifies and classifies new samples.

[0040] Further optimizing the technical solution, in step S8, a saliency segmentation model is constructed, and a segmentation mask is dynamically generated by performing multi-dimensional saliency calculation on the spectral response, texture complexity, and edge strength of the ultraviolet fluorescence image to achieve accurate positioning of the defect area;

[0041] Assume that the input ultraviolet fluorescence image is , generate a segmentation mask , defined as:

[0042] ;

[0043] in,

[0044] : saliency function, defined as the comprehensive saliency value of spectrum, texture and edge, used to evaluate whether a pixel belongs to a defect area;

[0045] : Adaptive threshold, dynamically calculated based on the global statistical properties of the saliency function;

[0046] : Output segmentation mask, where 1 represents the defect area and 0 represents the background area.

[0047] Further optimizing this technical solution, the significance function As shown below:

[0048] ;

[0049] in,

[0050] : Spectral significance, reflecting the abnormality of the textile surface in the ultraviolet fluorescence response;

[0051] : Texture saliency, describing the difference between texture complexity and surroundings;

[0052] : Edge saliency, used to capture local edge structure characteristics;

[0053] : Weight factor used to balance the importance of spectral, texture and edge saliency.

[0054] In a second aspect, an embodiment of the present invention provides a computer device comprising a memory and a processor, wherein the memory stores a computer program, wherein: when the computer program instructions are executed by the processor, the steps of a high-precision detection method for key points of textiles based on ultraviolet rays are implemented as described in the first aspect of the present invention.

[0055] In a third aspect, an embodiment of the present invention provides a computer-readable storage medium having a computer program stored thereon, wherein: when the computer program instructions are executed by a processor, the steps of a high-precision detection method for key points of textiles based on ultraviolet rays are implemented as described in the first aspect of the present invention.

[0056] Compared with the existing technology, the present invention provides a high-precision detection method for key points of textiles based on ultraviolet light, which has the following beneficial effects:

[0057] This UV-based high-precision textile keypoint detection method incorporates a spectral-texture joint distribution model and a saliency segmentation model to achieve deep fusion of multidimensional features and precise keypoint location. Supported by uniform UV illumination and combined with multidimensional saliency analysis, it significantly improves detection robustness and segmentation accuracy in complex backgrounds. This method is adaptable to textiles of various materials and craftsmanship, accurately detecting even minor defects and providing an efficient and stable solution for textile quality control. BRIEF DESCRIPTION OF THE DRAWINGS

[0058] In order to more clearly illustrate the technical solutions of the embodiments of the present invention, the following briefly introduces the drawings required for use in the description of the embodiments. Obviously, the drawings described below are only some embodiments of the present invention. For ordinary technicians in this field, other drawings can be obtained based on these drawings without paying any creative work.

[0059] Figure 1 This is a flow chart of a high-precision detection method for key points of textiles based on ultraviolet light proposed by the present invention;

[0060] Figure 2 This is a flow chart of the spectral-texture joint distribution model in the ultraviolet-based high-precision detection method for textile key points proposed by the present invention. DETAILED DESCRIPTION

[0061] In order to make the above-mentioned objects, features and advantages of the present invention more obvious and easy to understand, the specific embodiments of the present invention are described in detail below with reference to the accompanying drawings.

[0062] In the following description, many specific details are set forth to facilitate a full understanding of the present invention. However, the present invention may also be implemented in other ways different from those described herein. Those skilled in the art may make similar generalizations without violating the connotation of the present invention. Therefore, the present invention is not limited to the specific embodiments disclosed below.

[0063] Secondly, the term "one embodiment" or "embodiment" herein refers to a specific feature, structure, or characteristic that may be included in at least one implementation of the present invention. The phrase "in one embodiment" appearing in various places throughout this specification does not necessarily refer to the same embodiment, nor does it designate a separate or selective embodiment that is mutually exclusive with other embodiments.

[0064] Example 1:

[0065] Reference Figure 1 , which is the first embodiment of the present invention, provides a high-precision detection method for key points of textiles based on ultraviolet light, comprising the following steps:

[0066] S1. Determine the UV wavelength for detection based on fluorescence spectroscopy

[0067] In this embodiment, the initial phase of the detection method involves irradiating the textile surface with an ultraviolet (UV) light source to stimulate a fluorescence reaction. Different textiles or fabric treatments react differently to UV light, so a spectral analysis is first performed to evaluate the fluorescence response characteristics of the textile after UV exposure. By testing the effects of different UV wavelengths, the optimal UV wavelength can be determined and potential hidden defects or microcracks on the textile surface can be identified.

[0068] S2. Constructing a UV fluorescence spectral feature library for textiles

[0069] In this embodiment, ultraviolet irradiation experiments are performed on textile samples of different batches, materials, and dyeing methods, and their fluorescence spectrum characteristics are recorded to form an ultraviolet fluorescence spectrum characteristic library;

[0070] The feature library contains all variables that may affect detection accuracy, including factors such as textile material, dye type, and fabric structure. This data provides a scientific basis for subsequent feature matching and key point detection, and can improve the robustness and accuracy of key point detection in actual operations.

[0071] S3. Conduct a comprehensive analysis of the detection area and adjust the ultraviolet light source

[0072] In this embodiment, the intensity of ultraviolet radiation needs to be uniform to avoid data deviation caused by local strong or weak light. The ultraviolet intensity distribution distribution of the detection area is detected using an ultraviolet intensity distribution measuring instrument;

[0073] By adjusting the position, angle and irradiation distance of the UV light source, uniform UV distribution can be achieved, ensuring that each inspection point on the textile is evenly irradiated by UV rays, thereby avoiding detection errors caused by uneven UV intensity.

[0074] S4. Use a high-resolution UV camera to capture images of textile surfaces

[0075] In this example, the UV camera needs to have a high frame rate and high resolution to accurately capture subtle defects or structural changes that may exist on the textile surface. The UV camera should also be able to automatically adjust focus and exposure time to ensure clear and accurate image data under varying lighting conditions, reducing manual intervention and improving inspection efficiency and consistency.

[0076] S5. Preprocessing the collected ultraviolet fluorescence image

[0077] In this embodiment, the preprocessing includes the following steps:

[0078] First, an image filtering algorithm is applied to remove background noise and eliminate ambient light interference;

[0079] Then, an image enhancement algorithm is used to enhance the image contrast, especially in areas with weaker fluorescence, making tiny defects on the textile surface more obvious. This process ensures that the processed image does not lose real information due to excessive enhancement, thus ensuring image quality.

[0080] S6. Extract key point features from the preprocessed image

[0081] In this embodiment, when extracting key point features, a spectrum-texture joint distribution model is constructed to integrate the fluorescence response, texture changes, and shape characteristics of the textile surface, and use weight factors to couple them into a unified feature extraction framework;

[0082] In the spectral texture joint distribution model, let any pixel point on the textile surface be , its eigenvalue is calculated by the following formula:

[0083] ;

[0084] in,

[0085] : Pixels The comprehensive eigenvalue of is used to indicate the possibility of the point being a key point;

[0086] : The eigenvalue based on texture gradient is used to represent the texture complexity of the pixel;

[0087] : Spectral intensity characteristic value, calculated by the spectral response of the pixel point in the ultraviolet fluorescence image;

[0088] : Shape eigenvalue, calculated by the significance of the pixel in the surrounding geometric structure;

[0089] : Weight factor, indicating the importance of texture, spectrum, and shape features in key point extraction.

[0090] S7. Identify and classify defects using the extracted key point features

[0091] In this embodiment, the extracted key point features are input into a machine learning model. Using training data from a sample set, the model learns the characteristics of different types of textile defects and automatically identifies and classifies new samples. Common machine learning algorithms, such as support vector machines (SVMs) and convolutional neural networks (CNNs), can be selected based on the specific needs of the sample. This step can greatly improve the automation of defect identification and reduce errors caused by human intervention.

[0092] S8. Based on defect recognition, the defect area is accurately located through image segmentation

[0093] In this embodiment, a saliency segmentation model is constructed, and a segmentation mask is dynamically generated by performing multi-dimensional saliency calculation on the spectral response, texture complexity, and edge strength of the ultraviolet fluorescence image to achieve accurate positioning of the defect area.

[0094] Assume that the input ultraviolet fluorescence image is , generate a segmentation mask , defined as:

[0095] ;

[0096] in,

[0097] : saliency function, defined as the comprehensive saliency value of spectrum, texture and edge, used to evaluate whether a pixel belongs to a defect area;

[0098] : Adaptive threshold, dynamically calculated based on the global statistical properties of the saliency function;

[0099] : Output segmentation mask, where 1 represents the defect area and 0 represents the background area.

[0100] S9. Output test results and generate test report

[0101] In this example, the results of each step are integrated to generate a detailed defect report. This report not only includes the defect location and type, but also provides corresponding inspection parameters, treatment steps, and possible repair suggestions. This report provides an important basis for textile quality control and improvement, helping relevant personnel make subsequent decisions.

[0102] Example 2:

[0103] Reference Figure 2 , which is the second embodiment of the present invention, provides a method for calculating eigenvalues in a spectral-texture joint distribution model, the method comprising:

[0104] Texture gradient feature calculation

[0105] The texture changes of the textile surface are calculated by gradient. The complexity of the pixel point on the texture is calculated using the second-order derivative:

[0106] ;

[0107] in, represents the image intensity value, and The images are and This characteristic value can effectively distinguish areas with obvious structural changes on the textile surface, such as break points or discontinuous lines.

[0108] Spectral intensity feature calculation

[0109] The spectral intensity in the ultraviolet fluorescence image is used as the basis for determining key points. The spectral characteristic value is defined as:

[0110] ;

[0111] in, Yes In wavelength The fluorescence intensity under . It is the effective wavelength range of ultraviolet light (selected based on experiments, such as 250-400nm).

[0112] This feature reflects differences in the textile's properties under UV light, such as changes in spectral response caused by dye distribution or material defects.

[0113] Shape feature calculation

[0114] Define shape eigenvalues using the saliency of local geometric structure:

[0115] ;

[0116] in, Indicates a point The local geometric descriptor of θ (such as the response of a Gabor filter). Yes The N neighboring pixels around it.

[0117] The shape feature value calculates the significance by the geometric difference with the neighboring pixels, which is suitable for detecting cracks, holes or other deformations on the textile surface.

[0118] Weight optimization and feature fusion

[0119] Initial weight The performance of key point extraction can be optimized by cross-validation on actual data sets. In addition, an adaptive mechanism can be introduced:

[0120] ;

[0121] in, It is the variance of each feature dimension and is used to adaptively adjust the weights so that the model can dynamically adjust according to the characteristics of the input data.

[0122] This model integrates texture, spectral, and shape features to improve the accuracy of keypoint extraction. An adaptive weight optimization mechanism enhances the model's adaptability to different textile materials and dye types. The inclusion of shape features effectively detects complex cracks or local structural anomalies.

[0123] This embodiment also provides a method for calculating the saliency function. The saliency function in the saliency segmentation model is As shown below:

[0124] ;

[0125] in,

[0126] : Spectral significance, reflecting the abnormality of the textile surface in the ultraviolet fluorescence response;

[0127] : Texture saliency, describing the difference between texture complexity and surroundings;

[0128] : Edge saliency, used to capture local edge structure characteristics;

[0129] : Weight factor used to balance the importance of spectral, texture and edge saliency.

[0130] Based on this model, significant component calculations are performed, including:

[0131] Spectral significance calculation

[0132] Detecting defects on textile surfaces using spectral anomalies:

[0133] ;

[0134] in, Yes In wavelength The fluorescence intensity under .

[0135] It is the spectral mean of the pixel in its neighborhood, indicating a normal spectral distribution.

[0136] This significance value can effectively detect spectral anomalies such as uneven dye distribution or material changes.

[0137] Texture saliency calculation

[0138] Saliency is defined by comparing texture features with surrounding areas:

[0139] ;

[0140] in, Yes The texture feature values of (e.g., calculated by Gabor filter).

[0141] is the average texture feature value of the local neighborhood of the pixel.

[0142] High-value areas usually correspond to defective areas with abrupt changes in texture, such as break points or weaving gaps.

[0143] Edge saliency calculation

[0144] Compute saliency via gradient-boosted edge detection:

[0145] ;

[0146] Edge saliency reflects the changes in the local structure of the textile surface and helps to accurately locate boundaries such as cracks and holes.

[0147] When using this model, the following processes are involved:

[0148] Saliency map generation: For the input fluorescence image, the spectral saliency, texture saliency, and edge saliency are calculated separately, and then the comprehensive saliency map is calculated based on the saliency function.

[0149] Adaptive threshold segmentation: Based on the statistical characteristics of the saliency map, an adaptive threshold is calculated. For each pixel in the saliency map, its saliency value is determined to be greater than the threshold. If it is greater, it is classified as a defect area.

[0150] Post-processing optimization: Apply morphological operations (such as dilation and erosion) and connectivity analysis to the generated segmentation mask to eliminate noise points and fill small area breaks to obtain accurate defect segmentation results.

[0151] Example 3:

[0152] This embodiment also provides a computer device suitable for a high-precision detection method for key points of textiles based on ultraviolet rays, including a memory and a processor; the memory is used to store computer-executable instructions, and the processor is used to execute computer-executable instructions to implement the high-precision detection method for key points of textiles based on ultraviolet rays proposed in the above embodiment.

[0153] This embodiment further provides a storage medium having a computer program stored thereon. When the program is executed by a processor, the high-precision detection method for key points of textiles based on ultraviolet light as proposed in the above embodiment is implemented.

[0154] The computer device may be a terminal, comprising a processor, memory, a communication interface, a display, and an input device connected via a system bus. The processor provides computing and control capabilities. The memory includes non-volatile storage media and internal memory. The non-volatile storage media stores an operating system and computer programs. The internal memory provides an environment for the operating system and computer programs stored in the non-volatile storage media. The communication interface of the computer device is used to communicate with external terminals via wired or wireless communication. Wireless communication may be achieved via Wi-Fi, a carrier network, NFC (near-field communication), or other technologies. The display of the computer device may be a liquid crystal display or an electronic ink display. The input device may be a touchscreen overlay on the display, buttons, a trackball, or a touchpad on the computer device housing, or an external keyboard, touchpad, or mouse.

[0155] If a function is implemented as a software functional unit and sold or used as an independent product, it can be stored in a computer-readable storage medium. Based on this understanding, the technical solution of the present invention, or the portion that contributes to the prior art, or a portion of the technical solution, can be embodied in the form of a software product. This computer software product is stored in a storage medium and includes several instructions for causing a computer device (which can be a personal computer, server, or network device, etc.) to execute all or part of the steps of the various embodiments of the method of the present invention. The aforementioned storage medium includes various media that can store program code, such as USB flash drives, mobile hard drives, read-only memories (ROMs), random access memories (RAMs), magnetic disks, or optical disks.

[0156] The logic and / or steps represented in the flowcharts or otherwise described herein, for example, can be considered as a sequenced list of executable instructions for implementing the logical functions, and can be embodied in any computer-readable medium for use by, or in conjunction with, an instruction execution system, apparatus, or device (e.g., a computer-based system, a system including a processor, or other system that can fetch and execute instructions from an instruction execution system, apparatus, or device). For purposes of this specification, a "computer-readable medium" can be any device that can contain, store, communicate, propagate, or transport a program for use by, or in conjunction with, an instruction execution system, apparatus, or device.

[0157] More specific examples (a non-exhaustive list) of computer-readable media include the following: an electrical connection with one or more wires (electronic devices), a portable computer disk cartridge (magnetic devices), a random access memory (RAM), a read-only memory (ROM), an erasable and programmable read-only memory (EPROM or flash memory), a fiber optic device, and a portable compact disc read-only memory (CDROM). In addition, the computer-readable medium may even be paper or other suitable medium on which the program is printed, since the program may be obtained electronically, for example, by optically scanning the paper or other medium and then editing, interpreting, or processing it in another suitable manner as necessary, and then storing it in a computer memory.

[0158] It should be understood that various parts of the present invention can be implemented using hardware, software, firmware, or a combination thereof. In the above-described embodiments, multiple steps or methods can be implemented using software or firmware stored in a memory and executed by a suitable instruction execution system. For example, if implemented using hardware, as in another embodiment, any one of the following technologies known in the art or a combination thereof can be used: a discrete logic circuit having logic gate circuits for implementing logic functions on data signals, an application-specific integrated circuit having suitable combinational logic gate circuits, a programmable gate array (PGA), a field-programmable gate array (FPGA), etc.

[0159] It should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention and are not intended to limit the present invention. Although the present invention has been described in detail with reference to the preferred embodiments, those skilled in the art should understand that the technical solutions of the present invention may be modified or replaced by equivalents without departing from the spirit and scope of the technical solutions of the present invention, which should all be included in the scope of the claims of the present invention.

Claims

1. A high-precision detection method for key points of textiles based on ultraviolet light, characterized in that: The following steps are involved: S1. Determine the wavelength of ultraviolet light for detection based on fluorescence spectroscopy analysis; S2. Constructing a UV fluorescence spectrum feature library for textiles; S3. Conduct a comprehensive analysis of the detection area and adjust the UV light source; S4, using a high-resolution UV camera to capture images of the textile surface; S5, preprocessing the collected ultraviolet fluorescence image; S6. Extract key point features from the preprocessed image. During the key point feature extraction, a spectrum-texture joint distribution model is constructed to integrate the fluorescence response, texture change, and shape characteristics of the textile surface, and couple them using weight factors. In the spectral texture joint distribution model, let any pixel point on the textile surface be , the eigenvalue is calculated using the following formula: ； in, : Pixels The comprehensive eigenvalue of is used to indicate the possibility of the point being a key point; : The eigenvalue based on texture gradient is used to represent the texture complexity of the pixel point. The texture change of the textile surface is calculated by the gradient, and the complexity of the pixel point on the texture is calculated by the second-order derivative; : Spectral intensity characteristic value, calculated by the spectral response of the pixel point in the ultraviolet fluorescence image, and the spectral intensity in the ultraviolet fluorescence image is used as the basis for determining the key points; : Shape eigenvalue, calculated by the significance of the pixel in the surrounding geometric structure; : Weight factor, indicating the importance of texture, spectrum, and shape features in key point extraction; introducing an adaptive mechanism: ； in, It is the variance of each feature dimension, which is used to adaptively adjust the weights so that the model can dynamically adjust according to the characteristics of the input data; S7, identifying and classifying defects using the extracted key point features; S8. Based on defect recognition, the defect area is accurately located through image segmentation; S9. Output the test results and generate a test report.

2. The high-precision detection method for key points of textiles based on ultraviolet light according to claim 1, characterized in that: In step S1, the surface of the textile is irradiated with an ultraviolet light source to stimulate a fluorescence reaction on the surface of the textile, and a spectrum analysis is performed to evaluate the fluorescence response characteristics of the textile after ultraviolet irradiation, and to determine the ultraviolet wavelength for detection.

3. The high-precision detection method for key points of textiles based on ultraviolet light according to claim 1, characterized in that: In step S2, ultraviolet irradiation experiments are performed on textile samples of different batches, materials, and dyeing methods, and their fluorescence spectrum characteristics are recorded to form an ultraviolet fluorescence spectrum characteristic library; The feature library contains all variables that may affect the detection accuracy, including factors such as textile material, dye type, and fabric structure.

4. The high-precision detection method for key points of textiles based on ultraviolet light according to claim 1, characterized in that: In step S3, ultraviolet intensity distribution is measured in the detection area using an ultraviolet intensity distribution measuring instrument; By adjusting the position, angle and irradiation distance of the UV light source, uniform UV distribution can be achieved, ensuring that each inspection point on the textile is evenly irradiated by UV rays.

5. The high-precision detection method for key points of textiles based on ultraviolet light according to claim 1, characterized in that: In step S5, the pre-processing includes the following steps: First, an image filtering algorithm is applied to remove background noise and eliminate ambient light interference; Then, the image enhancement algorithm is used to enhance the contrast of the image and enhance the image of the areas with weak fluorescence, so that the tiny defects on the surface of the textile are clearly presented.

6. The high-precision detection method for key points of textiles based on ultraviolet light according to claim 5, characterized in that: In the spectral texture joint distribution model, the calculation of the eigenvalue includes: Texture gradient feature calculation: Spectral intensity feature calculation; Shape feature calculation; Weight optimization and feature fusion.

7. The high-precision detection method for key points of textiles based on ultraviolet light according to claim 1, characterized in that: In step S7, the extracted key point features are input into the machine learning model. Through the training set sample data, the machine learning model learns the characteristics of different types of textile defects and automatically identifies and classifies new samples.

8. The high-precision detection method for key points of textiles based on ultraviolet light according to claim 1, characterized in that: In step S8, a saliency segmentation model is constructed, and a segmentation mask is dynamically generated by performing multi-dimensional saliency calculation on the spectral response, texture complexity, and edge strength of the ultraviolet fluorescence image to achieve accurate positioning of the defect area; Assume that the input ultraviolet fluorescence image is , generate a segmentation mask , defined as: ； in, : saliency function, defined as the comprehensive saliency value of spectrum, texture and edge, used to evaluate whether a pixel belongs to a defect area; : Adaptive threshold, dynamically calculated based on the global statistical properties of the saliency function; : Output segmentation mask, where 1 represents the defect area and 0 represents the background area.

9. The high-precision detection method for key points of textiles based on ultraviolet light according to claim 8, characterized in that: The significance function As shown below: ； in, : Spectral significance, reflecting the abnormality of the textile surface in the ultraviolet fluorescence response; : Texture saliency, describing the difference between texture complexity and surroundings; : Edge saliency, used to capture local edge structure characteristics; : Weight factor used to balance the importance of spectral, texture and edge saliency.

Images