Textile fiber component intelligent detection method, system and equipment and storage medium

Abstract

The invention discloses an intelligent detection method, system and device for textile fiber components and a storage medium, and the method comprises the steps: extracting contour features based on a cross section image under the condition that the cross section image and Raman spectrum data of a to-be-detected textile are obtained; extracting spectrum sequence features based on the Raman spectrum data; according to the contour features and the spectral sequence features, determining a fiber classification result of the to-be-detected textile and a corresponding confidence threshold thereof through a trained fiber classification model; under the condition that the corresponding confidence threshold value is larger than a preset threshold value, the structure characteristic value of each fiber in the cross section image in the fiber classification result is determined, and the structure characteristic value comprises the average cross section area, the volume density and the number of the fibers; and the mass percentage content of each fiber in the fiber classification result is calculated according to the structure characteristic value, so that the efficiency and the accuracy of fiber type and mass percentage content detection are greatly improved.

Description

Technical Field

[0001] This application relates to the technical field of intelligent detection of textile fiber composition, and in particular to an intelligent detection method, system, device and storage medium for textile fiber composition. Background Technology

[0002] Textiles are defined as products made from raw materials such as yarn, thread, filament, rayon, chemical fibers, and metal wires, through spinning, weaving, and dyeing. With people's ever-growing needs for a better life, the materials used in textiles have become increasingly diverse, from traditional natural fibers to constantly evolving new fibers. This not only meets consumers' various requirements for the appearance of clothing but also improves the comfort and functionality of textiles. However, many merchants frequently use inferior or counterfeit products, posing potential health risks to consumers, disrupting market order, and infringing upon people's legitimate rights. Therefore, fiber testing has become one of the important items in textile testing.

[0003] Currently, the main methods for fiber detection include chemical dissolution, near-infrared spectroscopy, and microscopic counting. However, none of these methods can achieve both high detection efficiency and high quality in detecting fiber type and percentage by weight. Therefore, there is an urgent need for rapid and accurate detection of fiber type and percentage by weight in textiles. Summary of the Invention

[0004] This application aims to provide a method, system, device, and storage medium for intelligent detection of textile fiber composition, which can improve the efficiency and accuracy of fiber type and mass percentage detection.

[0005] In a first aspect, embodiments of this application provide a method for intelligent detection of textile fiber composition, comprising the following steps:

[0006] Having acquired a cross-sectional image and Raman spectral data of the textile to be inspected, contour features are extracted based on the cross-sectional image; spectral sequence features are extracted based on the Raman spectral data; wherein the cross-sectional image and the Raman spectral data have the same field of view;

[0007] Based on the contour features and the spectral sequence features, the fiber classification result of the textile to be detected and its corresponding confidence threshold are determined by the trained fiber classification model.

[0008] If the corresponding confidence threshold is greater than a preset threshold, the structural feature value of each fiber in the cross-sectional image is determined, wherein the structural feature value includes the average cross-sectional area, bulk density and fiber number;

[0009] The mass percentage of each fiber in the fiber classification results is calculated based on the structural feature values.

[0010] According to some embodiments of this application, the intelligent detection method for textile fiber composition further includes:

[0011] If the corresponding confidence threshold is less than the preset threshold, determine the roundness and irregularity of each fiber in the cross-sectional image in the fiber classification result;

[0012] Based on the circularity and the irregularity, the second fiber classification result of the textile to be tested is determined;

[0013] The second fiber classification result is taken as the final fiber classification result of the textile to be tested.

[0014] According to some embodiments of this application, determining the structural feature value of each fiber in the cross-sectional image in the fiber classification result includes:

[0015] The volume density, number of fibers, cross-sectional contour and number of pixels in the internal region of the contour for each fiber in the cross-sectional image are obtained from the fiber classification results, and the conversion factor of the actual physical size corresponding to each pixel in the image is obtained.

[0016] The cross-sectional area is obtained by multiplying the number of pixels by the conversion factor.

[0017] The cross-sectional areas of each fiber in the fiber classification results are added together to obtain the total cross-sectional area of ​​each fiber.

[0018] The average cross-sectional area is obtained by dividing the total cross-sectional area of ​​each fiber by the number of fibers. According to some embodiments of this application, determining the roundness of each fiber in the cross-sectional image in the fiber classification result includes:

[0019] Obtain the number of cross-sectional contour pixels of each fiber in the cross-sectional image and the conversion factor of the actual physical size corresponding to each pixel in the image in the fiber classification result;

[0020] Multiplying the cross-sectional area by four times pi yields the area product;

[0021] Multiply the number of pixels in the cross-sectional contour by the conversion factor to obtain the perimeter of the cross-section;

[0022] The circularity is obtained by dividing the product of the areas by the square of the cross-sectional perimeter.

[0023] According to some embodiments of this application, determining the degree of anisotropy of each fiber in the cross-sectional image in the fiber classification result includes:

[0024] Obtain the maximum inscribed circle diameter and minimum circumscribed circle diameter of each fiber in the cross-sectional image from the fiber classification results;

[0025] Calculate the ratio of the maximum inscribed circle diameter to the minimum circumscribed circle diameter;

[0026] Calculate the difference between one and the ratio, and use the difference as the degree of irregularity.

[0027] According to some embodiments of this application, the step of calculating the mass percentage of each fiber in the fiber classification result based on the structural feature value includes:

[0028] The product is obtained by multiplying the average cross-sectional area, bulk density and fiber count of the target fiber type in the fiber classification results.

[0029] The corresponding products of all fiber types in the fiber classification results are calculated and summed to obtain the total.

[0030] Divide the corresponding product by the sum to obtain the mass percentage of the target fiber type.

[0031] According to some embodiments of this application, the cross-sectional image is a transmission and reflection cross-sectional image, and the extraction of contour features based on the cross-sectional image and the extraction of spectral sequence features based on the Raman spectral data include:

[0032] Based on the transmissive and reflective cross-sectional image, the contour features are extracted using the YOLO algorithm;

[0033] Based on the Raman spectral data, the spectral sequence features are extracted using a preset spectral processing model, wherein the preset spectral processing model includes a CNN layer and a Transformer layer.

[0034] Secondly, embodiments of this application provide an intelligent detection system for textile fiber composition, the intelligent detection system for textile fiber composition comprising:

[0035] The data acquisition module is used to extract contour features based on the cross-sectional image and extract spectral sequence features based on the Raman spectral data, after acquiring the cross-sectional image and Raman spectral data of the textile to be inspected; wherein the cross-sectional image and the Raman spectral data have the same field of view;

[0036] The prediction module is used to determine the fiber classification result of the textile to be detected and its corresponding confidence threshold based on the contour features and the spectral sequence features, using a trained fiber classification model.

[0037] The structural feature value determination module is used to determine the structural feature value of each fiber in the cross-sectional image in the fiber classification result when the corresponding confidence threshold is greater than a preset threshold. The structural feature value includes average cross-sectional area, bulk density and fiber number.

[0038] The mass percentage calculation module is used to calculate the mass percentage of each fiber in the fiber classification result based on the structural feature value.

[0039] Thirdly, embodiments of this application provide an intelligent electronic device for detecting the fiber composition of textiles, including at least one control processor and a memory for communicatively connecting to the at least one control processor; the memory stores instructions executable by the at least one control processor, which are executed by the at least one control processor to enable the at least one control processor to perform the above-described intelligent detection method for the fiber composition of textiles.

[0040] Fourthly, embodiments of this application provide a computer-readable storage medium storing computer-executable instructions for causing a computer to execute the above-described intelligent detection method for textile fiber composition.

[0041] In this embodiment, after acquiring a cross-sectional image and Raman spectral data of the textile to be tested, contour features are extracted based on the cross-sectional image; spectral sequence features are extracted based on the Raman spectral data; wherein the field of view of the cross-sectional image and the Raman spectral data are the same; based on the contour features and spectral sequence features, the fiber classification result of the textile to be tested and its corresponding confidence threshold are determined through a trained fiber classification model; if the corresponding confidence threshold is greater than a preset threshold, the structural feature value of each fiber in the cross-sectional image in the fiber classification result is determined, wherein the structural feature value includes the average cross-sectional area, bulk density, and number of fibers; the mass percentage of each fiber in the fiber classification result is calculated based on the structural feature value, which greatly improves the efficiency and accuracy of fiber type and mass percentage detection.

[0042] Additional aspects and advantages of this application will be set forth in part in the description which follows, and in part will be obvious from the description, or may be learned by practice of this application. Attached Figure Description

[0043] The above and / or additional aspects and advantages of this application will become apparent and readily understood from the description of the embodiments taken in conjunction with the following drawings, in which:

[0044] Figure 1 This is a flowchart illustrating an embodiment of the intelligent detection method for textile fiber composition provided in this application;

[0045] Figure 2 This is a schematic diagram of the structure of an embodiment of the intelligent detection system for textile fiber composition provided in this application;

[0046] Figure 3 This is a schematic diagram of the structure of an embodiment of the electronic device provided in this application. Detailed Implementation

[0047] The embodiments of this application are described in detail below. Examples of these embodiments are shown in the accompanying drawings, wherein the same or similar reference numerals denote the same or similar elements or elements having the same or similar functions throughout. The embodiments described below with reference to the accompanying drawings are exemplary and are only used to explain this application, and should not be construed as limiting this application.

[0048] In the description of this application, the use of terms such as "first," "second," etc., is for the purpose of distinguishing technical features only and should not be construed as indicating or implying relative importance or implicitly indicating the number of technical features indicated or the order of the technical features indicated.

[0049] In the description of this application, it should be understood that the orientation descriptions, such as up, down, etc., are based on the orientation or positional relationship shown in the accompanying drawings, and are only for the convenience of describing this application and simplifying the description, and do not indicate or imply that the device or element referred to must have a specific orientation, or be constructed and operated in a specific orientation, and therefore should not be construed as a limitation of this application.

[0050] In the description of this application, it should be noted that, unless otherwise explicitly defined, terms such as "setup," "installation," and "connection" should be interpreted broadly, and those skilled in the art can reasonably determine the specific meaning of the above terms in this application in conjunction with the specific content of the technical solution.

[0051] Textiles are defined as products made from raw materials such as yarn, thread, filament, rayon, chemical fibers, and metal wires, through spinning, weaving, and dyeing. With people's ever-growing needs for a better life, the materials used in textiles have become increasingly diverse, from traditional natural fibers to constantly evolving new fibers. This not only meets consumers' various requirements for the appearance of clothing but also improves the comfort and functionality of textiles. However, many merchants frequently use inferior or counterfeit products, posing potential health risks to consumers, disrupting market order, and infringing upon people's legitimate rights. Therefore, fiber testing has become one of the important items in textile testing.

[0052] Currently, the main methods for fiber detection include chemical dissolution, near-infrared spectroscopy, and microscopic counting. However, none of these methods can achieve both high detection efficiency and high quality in detecting fiber type and percentage by weight. Therefore, there is an urgent need for rapid and accurate detection of fiber type and percentage by weight in textiles.

[0053] To address the aforementioned technical deficiencies, referring to... Figure 1 This method is applied to electronic devices, such as servers. Intelligent detection methods for textile fiber composition include:

[0054] Step S101: After obtaining the cross-sectional image and Raman spectral data of the textile to be detected, extract contour features based on the cross-sectional image; extract spectral sequence features based on the Raman spectral data; wherein the field of view of the cross-sectional image and the Raman spectral data are the same;

[0055] Step S102: Based on the contour features and spectral sequence features, determine the fiber classification result of the textile to be detected and its corresponding confidence threshold through the trained fiber classification model;

[0056] Step S103: When the corresponding confidence threshold is greater than the preset threshold, determine the structural feature value of each fiber in the cross-sectional image in the fiber classification result, wherein the structural feature value includes the average cross-sectional area, volume density and fiber number.

[0057] Step S104: Calculate the mass percentage of each fiber in the fiber classification results based on the structural characteristic values.

[0058] Based on contour features and spectral sequence features, the above-mentioned fiber classification results and corresponding confidence thresholds of the textile to be detected are determined through a trained fiber classification model, including:

[0059] The contour features and spectral sequence features are mapped to a common dimension by a fully connected layer, and gated fusion is performed along the channel dimension to obtain the fused features.

[0060] Based on the fused features, the fiber classification results of the textile to be detected and its corresponding confidence threshold are determined by the trained fiber classification model.

[0061] In this embodiment, after acquiring a cross-sectional image and Raman spectral data of the textile to be tested, contour features are extracted based on the cross-sectional image; spectral sequence features are extracted based on the Raman spectral data; wherein the field of view of the cross-sectional image and the Raman spectral data are the same; based on the contour features and spectral sequence features, the fiber classification result of the textile to be tested and its corresponding confidence threshold are determined through a trained fiber classification model; if the corresponding confidence threshold is greater than a preset threshold, the structural feature value of each fiber in the cross-sectional image in the fiber classification result is determined, wherein the structural feature value includes the average cross-sectional area, bulk density, and number of fibers; the mass percentage of each fiber in the fiber classification result is calculated based on the structural feature value, which greatly improves the efficiency and accuracy of fiber type and mass percentage detection.

[0062] In some embodiments, the intelligent detection method for textile fiber composition further includes:

[0063] If the corresponding confidence threshold is less than the preset threshold, determine the roundness and irregularity of each fiber in the cross-sectional image in the fiber classification results;

[0064] Based on roundness and irregularity, the second fiber classification result of the textile to be tested is determined;

[0065] The second fiber classification result is taken as the final fiber classification result of the textile to be tested.

[0066] The above-mentioned second fiber classification result of the textile to be tested can be determined manually based on roundness and irregularity.

[0067] This application determines the final fiber classification result by roundness and irregularity, which improves the traceability of the test results, supports a manual re-examination and correction mechanism, allows manual correction of inaccurately identified fiber outlines, and balances the efficiency of automation with the flexibility of manual calibration, thereby improving the accuracy of fiber type detection.

[0068] In some implementations, determining the structural feature values ​​of each fiber in the cross-sectional image in the fiber classification results includes:

[0069] Obtain the volume density, number of fibers, cross-sectional contour and number of pixels in the internal region of the contour for each fiber in the cross-sectional image, and the conversion factor of the actual physical size corresponding to each pixel in the image in the fiber classification results;

[0070] Multiply the number of pixels by the conversion factor to obtain the cross-sectional area.

[0071] Add up the cross-sectional areas of each fiber in the fiber classification results to get the total cross-sectional area of ​​each fiber.

[0072] The average cross-sectional area is obtained by dividing the total cross-sectional area of ​​each type of fiber by the number of fibers. This application improves the efficiency and accuracy of mass percentage content detection by calculating structural characteristic values ​​for subsequent mass percentage content calculation.

[0073] In some implementations, determining the roundness of each fiber in the cross-sectional image in the fiber classification results includes:

[0074] Obtain the number of cross-sectional contour pixels of each fiber in the cross-sectional image and the conversion factor of the actual physical size corresponding to each pixel in the image in the fiber classification results;

[0075] Multiplying the cross-sectional area by four times pi yields the area product;

[0076] Multiply the number of pixels in the cross-sectional outline by the conversion factor to obtain the perimeter of the cross-section;

[0077] The circularity is obtained by dividing the product of the areas by the square of the cross-sectional perimeter.

[0078] This application improves the accuracy of fiber type detection by calculating roundness to determine the final fiber classification result.

[0079] In some implementations, determining the degree of heterogeneity of each fiber in the cross-sectional image in the fiber classification results includes:

[0080] Obtain the maximum inscribed circle diameter and minimum circumscribed circle diameter of each fiber in the cross-sectional image from the fiber classification results;

[0081] Calculate the ratio of the largest inscribed circle diameter to the smallest circumscribed circle diameter;

[0082] Calculate the difference between 1 and the ratio, and use the difference as the degree of heterogeneity.

[0083] This application improves the accuracy of fiber type detection by calculating the degree of non-uniformity to determine the final fiber classification result.

[0084] In some implementations, the mass percentage of each fiber in the fiber classification results is calculated based on structural characteristic values, including:

[0085] The product is obtained by multiplying the average cross-sectional area, bulk density and fiber count of the target fiber type in the fiber classification results.

[0086] The sum is obtained by adding the corresponding products of all fiber types in the fiber classification results;

[0087] Divide the corresponding products by the sum to obtain the mass percentage of the target fiber type.

[0088] This application improves the efficiency and accuracy of mass percentage content detection by calculating the mass percentage content of the target fiber type based on the average cross-sectional area, bulk density, and fiber count.

[0089] In some implementations, the cross-sectional image is a transmission and reflection cross-sectional image, and contour features are extracted based on the cross-sectional image; spectral sequence features are extracted based on Raman spectral data, including:

[0090] Contour features are extracted from the transmissive and reflective cross-sectional images using the YOLO algorithm.

[0091] Based on Raman spectral data, spectral sequence features are extracted using a pre-defined spectral processing model, which includes a CNN layer and a Transformer layer.

[0092] Specifically, in some embodiments, based on Raman spectral data, the CNN layer first extracts local features from the fixed short-range wavenumber bands of the one-dimensional spectral sequence, and then concatenates them to form a new sequence. The Transformer layer then uses a self-attention mechanism to calculate the association weights between each position in the sequence and other positions, resulting in a weight matrix (recording the degree of association between each position and other positions). Then, the weights are used to sum the features of all positions for each position in the sequence to obtain another new sequence. Finally, the LinformerAttention mechanism mentioned below is used to reduce the dimensionality of the new sequence, and the spectral sequence features are obtained after pooling.

[0093] Specifically, in some embodiments, an image model is invoked to automatically identify and mark the cross-sectional contours of fibers, extract image features (flattened to 25600 dimensions), calculate the cross-sectional area of ​​each fiber, and statistically record data such as the number of fibers (n1, n2), average area (S1, S2), standard deviation, and coefficient of variation. The Raman spectroscopy processing module performs baseline correction (polynomial fitting), wavenumber axis alignment, noise reduction (SG filtering algorithm), and normalization on the acquired raw spectra, and invokes the spectral model to extract spectral sequence features (sequence length 3000, 128 dimensions) from the preprocessed spectral data. The image features (25600 dimensions) are reduced to 128 dimensions through a fully connected layer, and then fused with the spectral sequence features through gating to form a 256-dimensional fused feature vector, which is then input into the qualitative classification module. The qualitative classification module outputs a 95% confidence level for classifying a certain type of fiber as "cotton" and a 97% confidence level for classifying another type of fiber as "polyester." Since both confidence levels are ≥95%, the product is directly classified as "cotton" and "polyester," and thus identified as a "cotton / polyester blended product." The quantitative analysis module calculates the percentage content of each fiber type based on the data transmitted from the image model, outputting the corresponding percentage: cotton:polyester = 62.5%:37.5%. After the qualitative classification and quantitative analysis are completed, the multimodal intelligent recognition system automatically generates a test report.

[0094] Specifically, in some embodiments, when the qualitative classification module outputs a prediction confidence level of 95% for a certain type of fiber identified as "wool," which is greater than or equal to the set 95% threshold, and outputs a prediction confidence level of 56% for another type of fiber identified as "modal," which is less than 70%, it is determined to be "unknown component." Ultimately, the product is classified as both "wool" and "unknown component," and the blended product is identified as a "wool / unknown component blended product." The quantitative analysis module calculates the percentage content of each fiber type based on the data transmitted from the image model and outputs the corresponding percentage: wool:unknown component = 46.4%:53.6%. After the qualitative classification and quantitative analysis are completed, the multimodal intelligent recognition system automatically generates a detection report.

[0095] Specifically, in some embodiments, the manual re-inspection interface is opened, the image and spectrum are checked, and the unmarked fibers (2 fibers) are processed. The inspector manually marks them as "wool" based on the image and spectrum. After correction, the final mass percentage result of the fiber type is: wool: unknown component = 46.7%: 53.3%.

[0096] Additionally, refer to Figure 2 One embodiment of this application provides an intelligent detection system for textile fiber composition, including a data acquisition module 1100, a prediction module 1200, a structural feature value determination module 1300, and a mass percentage content calculation module 1400, wherein:

[0097] The data acquisition module 1100 is used to extract contour features based on the cross-sectional image and extract spectral sequence features based on the Raman spectral data when the cross-sectional image and Raman spectral data of the textile to be inspected are acquired; wherein the field of view of the cross-sectional image and the Raman spectral data are the same.

[0098] The prediction module 1200 is used to determine the fiber classification result of the textile to be detected and its corresponding confidence threshold based on the contour features and spectral sequence features, through a trained fiber classification model.

[0099] The structural feature value determination module 1300 is used to determine the structural feature value of each fiber in the cross-sectional image in the fiber classification result when the corresponding confidence threshold is greater than the preset threshold. The structural feature value includes the average cross-sectional area, volume density and fiber number.

[0100] The mass percentage calculation module 1400 is used to calculate the mass percentage of each fiber in the fiber classification results based on structural feature values.

[0101] It should be noted that the system embodiments described above are based on the same inventive concept as the method embodiments described above. Therefore, the relevant content of the method embodiments described above is also applicable to the system embodiments described above, and will not be repeated here.

[0102] Figure 3 A schematic diagram of the hardware structure for intelligent detection of textile fiber composition provided in an embodiment of this application is shown.

[0103] The intelligent detection device for textile fiber composition may include a processor 301 and a memory 302 storing computer program instructions.

[0104] Specifically, the processor 301 may include a central processing unit (CPU), an application-specific integrated circuit (ASIC), or one or more integrated circuits that can be configured to implement the embodiments of this application.

[0105] Memory 302 may include mass storage for data or instructions. For example, and not limitingly, memory 302 may include a hard disk drive (HDD), floppy disk drive, flash memory, optical disk, magneto-optical disk, magnetic tape, or Universal Serial Bus (USB) drive, or a combination of two or more of these. Where appropriate, memory 302 may include removable or non-removable (or fixed) media. Where appropriate, memory 302 may be internal or external to the integrated gateway disaster recovery device. In a particular embodiment, memory 302 is non-volatile solid-state memory.

[0106] In some embodiments, memory 302 may include read-only memory (ROM), random access memory (RAM), disk storage media device, optical storage media device, flash memory device, electrical, optical, or other physical / tangible memory storage device. Thus, generally, memory includes one or more tangible (non-transitory) computer-readable storage media (e.g., memory devices) encoded with software including computer-executable instructions, and when the software is executed (e.g., by one or more processors), it is operable to perform the operations described with reference to the method according to one aspect of this disclosure.

[0107] The processor 301 reads and executes computer program instructions stored in the memory 302 to implement any of the intelligent detection methods for textile fiber composition in the above embodiments.

[0108] In one example, the intelligent textile fiber composition detection device may also include a communication interface 303 and a bus 310. For example, Figure 3 As shown, the processor 301, memory 302, and communication interface 303 are connected through bus 310 and complete communication with each other.

[0109] The communication interface 303 is mainly used to realize communication between various modules, devices, units and / or equipment in the embodiments of this application.

[0110] Bus 310 includes hardware, software, or both, that couples components of a smart textile fiber composition detection device together. For example, and not limitingly, the bus may include an Accelerated Graphics Port (AGP) or other graphics bus, an Enhanced Industry Standard Architecture (EISA) bus, a Front Side Bus (FSB), HyperTransport (HT) interconnect, an Industry Standard Architecture (ISA) bus, an Infinite Bandwidth Interconnect, a Low Pin Count (LPC) bus, a memory bus, a Microchannel Architecture (MCA) bus, a Peripheral Component Interconnect (PCI) bus, a PCI-Express (PCI-X) bus, a Serial Advanced Technology Attachment (SATA) bus, a Video Electronics Standards Association Local (VLB) bus, or other suitable buses, or combinations of two or more of these. Where appropriate, bus 310 may include one or more buses. Although specific buses are described and illustrated in embodiments of this application, any suitable bus or interconnect is contemplated herein.

[0111] This intelligent textile fiber composition detection device can execute the intelligent textile fiber composition detection method described in this application embodiment based on a three-dimensional design model, thereby achieving a combination of... Figure 1 and Figure 2 The invention describes a method and system for intelligent detection of textile fiber composition.

[0112] Furthermore, in conjunction with the intelligent detection method for textile fiber composition in the above embodiments, this application embodiment can provide a computer storage medium for implementation. This computer storage medium stores computer program instructions; when these computer program instructions are executed by a processor, they implement any of the intelligent detection methods for textile fiber composition in the above embodiments.

[0113] It should be clarified that this application is not limited to the specific configurations and processes described above and shown in the figures. For the sake of brevity, detailed descriptions of known methods are omitted here. In the above embodiments, several specific steps are described and shown as examples. However, the method process of this application is not limited to the specific steps described and shown. Those skilled in the art can make various changes, modifications, and additions, or change the order of steps, after understanding the spirit of this application.

[0114] The functional blocks shown in the above block diagram can be implemented as hardware, software, firmware, or a combination thereof. When implemented in hardware, they can be, for example, electronic circuits, application-specific integrated circuits (ASICs), appropriate firmware, plug-ins, function cards, etc. When implemented in software, the elements of this application are programs or code segments used to perform the required tasks. Programs or code segments can be stored on a machine-readable medium or transmitted over a transmission medium or communication link via data signals carried on a carrier wave. "Machine-readable medium" can include any medium capable of storing or transmitting information. Examples of machine-readable media include electronic circuits, semiconductor memory devices, ROM, flash memory, erasable ROM (EROM), floppy disks, CD-ROMs, optical disks, hard disks, fiber optic media, radio frequency (RF) links, etc. Code segments can be downloaded via computer networks such as the Internet, intranets, etc.

[0115] It should also be noted that the exemplary embodiments mentioned in this application describe methods or systems based on a series of steps or apparatus. However, this application is not limited to the order of the above steps; that is, the steps can be performed in the order mentioned in the embodiments, or in a different order, or several steps can be performed simultaneously.

[0116] The aspects of this disclosure have been described above with reference to flowchart illustrations and / or block diagrams of methods, apparatus (systems), and computer program products according to embodiments of this disclosure. It should be understood that each block in the flowchart illustrations and / or block diagrams, and combinations of blocks in the flowchart illustrations and / or block diagrams, can be implemented by computer program instructions. These computer program instructions can be provided to a processor of a general-purpose computer, a special-purpose computer, or other programmable data processing apparatus to produce a machine such that these instructions, executable via the processor of the computer or other programmable data processing apparatus, enable the implementation of the functions / actions specified in one or more blocks of the flowchart illustrations and / or block diagrams. Such a processor can be, but is not limited to, a general-purpose processor, a special-purpose processor, a special application processor, or a field-programmable logic circuit. It is also understood that each block in the block diagrams and / or flowcharts, and combinations of blocks in the block diagrams and / or flowcharts, can also be implemented by special-purpose hardware performing the specified functions or actions, or can be implemented by a combination of special-purpose hardware and computer instructions.

[0117] The above description is merely a specific implementation of this application. Those skilled in the art will clearly understand that, for the sake of convenience and brevity, the specific working processes of the systems, modules, and units described above can be referred to the corresponding processes in the foregoing method embodiments, and will not be repeated here. It should be understood that the protection scope of this application is not limited thereto. Any person skilled in the art can easily conceive of various equivalent modifications or substitutions within the technical scope disclosed in this application, and these modifications or substitutions should all be covered within the protection scope of this application.

Claims

1. A method for intelligent detection of textile fiber composition, characterized in that, The intelligent detection method for textile fiber composition includes: Having acquired a cross-sectional image and Raman spectral data of the textile to be inspected, contour features are extracted based on the cross-sectional image; spectral sequence features are extracted based on the Raman spectral data; wherein the cross-sectional image and the Raman spectral data have the same field of view; Based on the contour features and the spectral sequence features, the fiber classification result of the textile to be detected and its corresponding confidence threshold are determined by the trained fiber classification model. If the corresponding confidence threshold is greater than a preset threshold, the structural feature value of each fiber in the cross-sectional image is determined, wherein the structural feature value includes the average cross-sectional area, bulk density and fiber number; The mass percentage of each fiber in the fiber classification results is calculated based on the structural feature values.

2. The method according to claim 1, characterized in that, The intelligent detection method for textile fiber composition also includes: If the corresponding confidence threshold is less than the preset threshold, determine the roundness and irregularity of each fiber in the cross-sectional image in the fiber classification result; Based on the circularity and the irregularity, the second fiber classification result of the textile to be tested is determined; The second fiber classification result is taken as the final fiber classification result of the textile to be tested.

3. The method according to claim 1, characterized in that, Determining the structural feature value of each fiber in the cross-sectional image in the fiber classification result includes: The volume density, number of fibers, cross-sectional contour and number of pixels in the internal region of the contour for each fiber in the cross-sectional image are obtained from the fiber classification results, and the conversion factor of the actual physical size corresponding to each pixel in the image is obtained. The cross-sectional area is obtained by multiplying the number of pixels by the conversion factor. The cross-sectional areas of each fiber in the fiber classification results are added together to obtain the total cross-sectional area of ​​each fiber. The average cross-sectional area is obtained by dividing the total cross-sectional area of ​​each fiber by the number of fibers.

4. The method according to claim 2, characterized in that, Determining the roundness of each fiber in the cross-sectional image in the fiber classification results includes: Obtain the number of cross-sectional contour pixels of each fiber in the cross-sectional image and the conversion factor of the actual physical size corresponding to each pixel in the image in the fiber classification result; Multiplying the cross-sectional area by four times pi yields the area product; Multiply the number of pixels in the cross-sectional contour by the conversion factor to obtain the perimeter of the cross-section; The circularity is obtained by dividing the product of the areas by the square of the cross-sectional perimeter.

5. The method according to claim 2, characterized in that, Determining the degree of heterogeneity of each fiber in the cross-sectional image in the fiber classification results includes: Obtain the maximum inscribed circle diameter and minimum circumscribed circle diameter of each fiber in the cross-sectional image from the fiber classification results; Calculate the ratio of the maximum inscribed circle diameter to the minimum circumscribed circle diameter; Calculate the difference between one and the ratio, and use the difference as the degree of irregularity.

6. The method according to claim 1, characterized in that, The calculation of the mass percentage of each fiber in the fiber classification result based on the structural feature value includes: The product is obtained by multiplying the average cross-sectional area, bulk density and fiber count of the target fiber type in the fiber classification results. The corresponding products of all fiber types in the fiber classification results are calculated and summed to obtain the total. Divide the corresponding product by the sum to obtain the mass percentage of the target fiber type.

7. The method according to claim 1, characterized in that, The cross-sectional image is a transmissive and reflective cross-sectional image, and the contour features are extracted based on the cross-sectional image; Extracting spectral sequence features based on the Raman spectral data includes: Based on the transmissive and reflective cross-sectional image, the contour features are extracted using the YOLO algorithm; Based on the Raman spectral data, the spectral sequence features are extracted using a preset spectral processing model, wherein the preset spectral processing model includes a CNN layer and a Transformer layer.

8. A smart detection system for textile fiber composition, characterized in that, The intelligent detection system for textile fiber composition includes: The data acquisition module is used to extract contour features based on the cross-sectional image and extract spectral sequence features based on the Raman spectral data, after acquiring the cross-sectional image and Raman spectral data of the textile to be inspected; wherein the cross-sectional image and the Raman spectral data have the same field of view; The prediction module is used to determine the fiber classification result of the textile to be detected and its corresponding confidence threshold based on the contour features and the spectral sequence features, using a trained fiber classification model. The structural feature value determination module is used to determine the structural feature value of each fiber in the cross-sectional image in the fiber classification result when the corresponding confidence threshold is greater than a preset threshold. The structural feature value includes average cross-sectional area, bulk density and fiber number. The mass percentage calculation module is used to calculate the mass percentage of each fiber in the fiber classification result based on the structural feature value.

9. A smart detection device for textile fiber composition, characterized in that, It includes at least one control processor and a memory for communicatively connecting to the at least one control processor; the memory stores instructions executable by the at least one control processor, which, when executed by the at least one control processor, enable the at least one control processor to perform a smart detection method for textile fiber composition as described in any one of claims 1 to 7.

10. A computer-readable storage medium, characterized in that: The computer-readable storage medium stores computer-executable instructions for causing a computer to perform a smart detection method for textile fiber composition as described in any one of claims 1 to 7.

Images