A method, system and electronic device for intelligent cloth inspection

Abstract

The application provides a kind of intelligent cloth inspection method, system and electronic equipment, it is related to intelligent detection field;Method includes: the bar code of the fabric roll to be inspected is scanned, the fabric roll specification is determined, so as to display the defect entry template corresponding to fabric specification on intelligent equipment;Obtain multiple image pictures of the part of fabric spread on the cloth inspection area under different light environments;After adopting cloth defect detection model to identify the fabric defect point set in the defect entry template, the fabric defect point in the image picture is obtained, and the score of fabric defect point is automatically judged according to defect scoring standard;According to fabric evaluation standard, each fabric defect point set and all fabric defect point deduction score, generate fabric score report;The fabric roll that completes cloth inspection is classified by grade.The application combines the use of cloth inspection machine and intelligent equipment, solves the technical problems of low efficiency and error in manual recording of fabric defect during cloth inspection, and can directly generate the score report of fabric roll for inventory management.

Description

Technical Field

[0001] This invention relates to the field of intelligent detection technology, specifically to an intelligent fabric inspection method, system, and electronic device. Background Technology

[0002] During operation, the fabric inspection machine provides a suitable hardware environment for continuous, segmented unfolding of the fabric, with ample light to assist operators in visually identifying surface defects and color differences. Simultaneously, it automatically completes length recording and roll-up sorting. Surface defects include black spots, hairs, foreign objects, insects, double weft threads, dense or sparse weft threads, broken warp threads, broken yarns, holes, and other surface imperfections. By utilizing the fabric inspection machine in conjunction with operators, the machine enables quality management of fabrics from manufacturers, allowing for the selection of high-quality fabrics in garment production and ultimately improving garment quality.

[0003] However, traditional defect detection methods relying on visual inspection by operators suffer from drawbacks such as fatigue during manual inspection and a high rate of missed detections. This not only fails to meet the increasingly stringent quality requirements of the apparel industry for qualified fabrics, but also results in low efficiency and an inability to quickly generate inspection reports due to the manual recording of defects during factory fabric inspections. Furthermore, manual inspection methods cannot be integrated with the factory's warehousing and operational systems in real time, hindering the rapid detection and resolution of anomalies when production problems arise. Summary of the Invention

[0004] The purpose of this invention is to provide a method, system, and electronic device for intelligent fabric inspection. By combining traditional fabric inspection machines with intelligent devices to systematically input fabric defects, an inspection report can be quickly generated. This not only improves input efficiency but also saves the time cost of manually re-entering data into the system and avoids input errors.

[0005] To achieve the above objectives, the present invention proposes the following technical solution:

[0006] Firstly, a method for intelligent fabric inspection is proposed, including:

[0007] Scan the barcode of the fabric roll to be inspected on the fabric inspection machine to determine the fabric roll specifications so that the corresponding fabric specification defect entry template can be displayed on the smart device.

[0008] Under different lighting conditions, multiple images of the fabric roll laid out on the upper part of the fabric inspection area were acquired sequentially.

[0009] The fabric defect detection model based on deep learning is used to identify fabric defects in each image picture in turn, and after merging into a fabric defect set of the corresponding part of the fabric, the fabric defects are recorded in the corresponding template area of the defect recording template one by one, and the fabric defect score is automatically judged according to the defect scoring standard; wherein the fabric defect detection model based on deep learning uses the Faster R-CNN algorithm of semi-supervised learning;

[0010] A defect recording end signal is received, and a fabric score report is generated according to the fabric evaluation standard, each fabric defect set, and the score of all fabric defects automatically judged according to the defect scoring standard; wherein the fabric score report lists the total score of the fabric roll, the number of different fabric defects, and the total score of each classified fabric defect;

[0011] According to the fabric grade evaluation standard and the fabric score report, the fabric roll that has completed the inspection is classified by grade.

[0012] Further, it also includes:

[0013] According to the size of the fabric roll and the size of the inspection area of the inspection machine, a positioning template matching the inspection area is set, and the positioning template includes a plurality of positioning areas;

[0014] For each fabric roll specification, a defect recording template corresponding to the positioning template of the corresponding fabric roll specification is configured on the intelligent device, and the defect recording template sets a plurality of defect positioning areas;

[0015] The image pictures are established with positioning coordinates, so that when the fabric defect detection model based on deep learning identifies and obtains the fabric defects, the first position information of the fabric defects is located synchronously, and the second position information of the fabric defects on the positioning template is determined on the inspection machine; wherein the first position information is the coordinate information in the positioning coordinates where the fabric defects are located, and the second position information is the positioning area where the fabric defects are located;

[0016] The position deviation of the first position information and the second position information is judged, and when the position deviation is within a preset range, the corresponding defect recording template is selected according to the specification of the fabric roll, the first position information or the second position information, and the fabric defects are recorded in the corresponding position of the defect recording template.

[0017] Further, it also includes:

[0018] According to the defect scoring standard, the defect recording template is divided into a plurality of scoring zones, and the scoring zones include a plurality of defect positioning areas;

[0019] According to the deduction score of each fabric defect, a corresponding deduction score item is added to each of the score zones, and the deduction score item includes a deduction score corresponding to different fabric defects; wherein the deduction score items of each defect positioning area in each of the score zones are the same.

[0020] Further, it further comprises:

[0021] According to the fabric roll specification, the length of the fabric wound thereon is determined;

[0022] According to the length of the fabric, the defect entry template corresponding to the fabric roll specification is divided into a page-turning text, any page of the page-turning text is similar to the fabric inspection area size, and each page of the page-turning text corresponds to each fabric section photographed by the fabric roll;

[0023] When the fabric in the fabric inspection area of the fabric inspection machine is detected, the fabric of the fabric roll is sequentially released to re-cover the fabric inspection area, and the corresponding defect entry template automatically switches to the next page after completing the defect entry of the current page.

[0024] Further, it further comprises:

[0025] When the fabric inspection machine is working, it is judged in real time whether the fabric roll in the current inspection belongs to a qualified fabric roll, and when the fabric roll belongs to a qualified fabric roll, the fabric roll is classified according to the fabric grade evaluation standard and the fabric score report;

[0026] Wherein, the process of judging in real time whether the fabric roll in the current inspection belongs to a qualified fabric roll comprises:

[0027] The fabric score of the fabric roll currently inspected by the fabric inspection machine is calculated in real time, and when the fabric score of the fabric roll being inspected in each preset inspection period is lower than the corresponding score threshold, the fabric roll is unqualified;

[0028] The types, quantities and corresponding areas of the entered fabric defects are calculated in real time, and when the number of target fabric defects exceeds a preset target or the area exceeds a preset proportion, the fabric roll is unqualified.

[0029] In the second aspect, an intelligent fabric inspection system is provided, comprising:

[0030] A scanning determination module is configured to scan the bar code of the fabric roll to be inspected on the fabric inspection machine to determine the fabric roll specification, so as to display the defect entry template corresponding to the fabric specification on the intelligent device;

[0031] An identification acquisition module is configured to sequentially acquire a plurality of image pictures of part of the fabric spread on the fabric inspection area under different light environments;

[0032] The input judgment module is configured to: adopt a fabric defect detection model based on deep learning to sequentially identify fabric defects in each image, merge the fabric defects into fabric defect sets corresponding to partial fabrics, and then input the fabric defects into corresponding template regions of the defect input template one by one, and automatically judge the fabric defects according to a defect scoring standard to obtain fabric defect deduction scores; wherein the fabric defect detection model based on deep learning adopts a Faster R-CNN algorithm based on semi-supervised learning.

[0033] The receiving and generating module is configured to receive a defect input end signal, generate a fabric score report according to a fabric evaluation standard, fabric defect sets, and all fabric defect deduction scores automatically judged according to the defect scoring standard; wherein the fabric score report lists total fabric deduction scores, numbers of different fabric defects, and total deduction scores of classified fabric defects.

[0034] The classification module is configured to classify the fabric rolls according to a fabric grade evaluation standard and the fabric score report.

[0035] Further, the method further comprises:

[0036] The setting module is configured to set a positioning template matched with a fabric inspection area according to a size of the fabric roll and a size of the fabric inspection area of the fabric inspection machine, wherein the positioning template comprises a plurality of positioning regions.

[0037] The configuration module is configured to configure, for each fabric roll, a defect input template corresponding to the positioning template of the corresponding fabric roll on the intelligent device, wherein the defect input template comprises a plurality of defect positioning regions.

[0038] The obtaining and determining module is configured to establish a positioning coordinate for the image, so that when the fabric defect detection model based on deep learning identifies and obtains the fabric defects, the first position information of the fabric defects is located synchronously, and the second position information of the fabric defects on the positioning template of the fabric inspection machine is determined; wherein the first position information is coordinate information in a positioning coordinate where the fabric defects are located, and the second position information is a positioning region where the fabric defects are located.

[0039] The judgment and input module is configured to judge a position deviation of the first position information and the second position information, and when the position deviation is within a preset range, select the corresponding defect input template according to the size of the fabric roll, the first position information, or the second position information, and input the fabric defects at the corresponding position of the defect input template.

[0040] Further, the method further comprises:

[0041] The division unit is configured to divide the defect input template into a plurality of scoring zones according to a defect scoring standard, wherein each scoring zone comprises a plurality of defect positioning regions.

[0042] An additional unit is configured to add corresponding deduction score items to each of the score zones according to the deduction scores of each fabric defect, and the deduction score items include deduction scores corresponding to different fabric defects.

[0043] Further, the method further comprises:

[0044] A judging module is configured to judge whether the fabric roll in current inspection belongs to a qualified fabric roll when the cloth inspection machine is working, and perform grade division on the fabric roll according to the fabric grade evaluation standard and the fabric score report when the fabric roll belongs to a qualified fabric roll.

[0045] The process of judging whether the fabric roll in current inspection belongs to a qualified fabric roll in real time comprises:

[0046] The fabric score of the fabric roll currently inspected by the cloth inspection machine is calculated in real time, and the fabric roll is unqualified when the fabric score of the fabric roll in each preset inspection period is lower than the corresponding score threshold.

[0047] The type, number and corresponding area of the input fabric defects are calculated in real time, and the fabric roll is unqualified when the number of target fabric defects exceeds a preset target or the area exceeds a preset proportion.

[0048] In a third aspect, an electronic device is provided, which includes a computer program stored in a computer readable storage medium; when a processor of the electronic device reads the computer program from the computer readable storage medium, the processor executes the computer program, so that the electronic device executes the steps of the intelligent cloth inspection method described above.

[0049] According to the above technical solutions, the technical solutions of the present application have the following beneficial effects:

[0050] The disclosed intelligent cloth inspection method, system and electronic device, the method comprising: scanning the bar code of the fabric roll to be inspected on the cloth inspection machine, determining the fabric roll specification, so as to display the defect entry template corresponding to the fabric specification on the intelligent device; sequentially acquiring multiple image pictures of the part of fabric spread on the cloth inspection area under different light environments; using a cloth defect detection model based on deep learning to identify the fabric defects in each image picture in turn, merging the fabric defect sets corresponding to the part of fabric, and then entering the fabric defects one by one in the corresponding template area of the defect entry template, and automatically judging the fabric defect deduction score according to the defect scoring standard; receiving the defect entry end signal, generating a fabric score report according to the fabric evaluation standard, each fabric defect set and the automatically judged fabric defect deduction score according to the defect scoring standard; and classifying the fabric roll that has completed the cloth inspection according to the fabric grade evaluation standard and the fabric score report. The scheme combines the cloth inspection machine and the intelligent device to solve the technical problems of low efficiency of manual observation and manual recording of fabric defects during cloth inspection, and errors in secondary entry.

[0051] When the fabric defect electronic entry is performed, the scheme realizes the rapid positioning of fabric defects on fabric rolls of different specifications and the deduction score judgment through the positioning template arranged on the cloth inspection machine and the defect entry template arranged on the intelligent device. The deduction score item attached to each defect positioning area can quickly generate a score report to judge whether the fabric roll currently inspected is qualified and the grade of the fabric roll. Meanwhile, the intelligent device can be used to compare the quality of multiple fabric rolls, and after the corresponding relationship between the cloth inspection data and the bar code is stored, the intelligent device can be connected to the WMS warehouse system and the factory ERP business system to perform inventory management and assist in business promotion. In addition, the rapid discovery of problem fabric rolls helps to discover and solve problems in time, and helps to improve production efficiency and production quality.

[0052] It should be understood that all combinations of the aforementioned concepts and additional concepts described in greater detail below can be seen as part of the subject matter of the present disclosure provided such concepts are not mutually inconsistent.

[0053] The foregoing and other aspects, embodiments and features of the present teachings can be better understood from the following description of the present teachings taken in conjunction with the accompanying drawings. Additional aspects, embodiments and features of the present teachings will be apparent from the following description {e.g., examples of exemplary embodiments) taken in conjunction with the accompanying drawings. BRIEF DESCRIPTION OF DRAWINGS

[0054] The accompanying drawings are not necessarily drawn to scale. In the drawings, each identical, or nearly identical, component that is illustrated in various figures is represented with a like numeral. For purposes of clarity, not every component is called out in every drawing. There is now being described, by way of example, embodiments of various aspects of the present application with reference to the accompanying drawings, in which:

[0055] Figure 1 Method flow chart of intelligent cloth inspection disclosed for embodiments of the present application;

[0056] Figure 2 Fabric defect entry flow chart according to defect entry template disclosed for embodiments of the present application;

[0057] Figure 3 Process flow chart of adding score deduction items to defect entry template disclosed for embodiments of the present application;

[0058] Figure 4 Configuration defect entry template example chart disclosed for embodiments of the present application;

[0059] Figure 5 Fabric roll classification flow chart disclosed for embodiments of the present application;

[0060] Figure 6 System structure block chart of intelligent cloth inspection disclosed for embodiments of the present application;

[0061] Figure 7 Structure block chart of electronic device disclosed for embodiments of the present application. DETAILED DESCRIPTION

[0062] In order to make the objects, technical solutions and advantages of the embodiments of the present application clearer, the technical solutions of the embodiments of the present application will be described clearly and completely below with reference to the drawings of the embodiments of the present application. Obviously, the described embodiments are some, but not all, of the embodiments of the present application. Based on the described embodiments of the present application, all other embodiments obtained by those of ordinary skill in the art without any creative work fall within the scope of protection of the present application. Unless otherwise defined, the technical terms or scientific terms used herein should be understood as the common meanings thereof by those of ordinary skill in the art.

[0063] The terms "first", "second", and similar terms used in the specification and claims of the present patent application do not denote any order, quantity, or importance, but are used to distinguish different constituent parts. Also, the singular forms "a", "an", and "the" do not denote the quantity limitation, but denote the existence of at least one, unless the context clearly indicates otherwise. The terms "comprise", "comprising", and similar terms mean that the elements or objects appearing before "comprise" or "comprising" encompass the features, integers, steps, operations, elements, and / or components listed after "comprise" or "comprising", and do not exclude the presence or addition of one or more other features, integers, steps, operations, elements, components, and / or groups thereof.

[0064] Based on the conventional defect detection scheme using the visual method of the operator, the operator usually records the fabric defects manually. Manual recording is not only inefficient, but also cannot quickly evaluate and generate the fabric inspection report, and the manual recording materials are easy to be lost, so secondary input into the electronic system is often required, and recording errors are easy to occur during the secondary input. Therefore, the present application aims to provide a smart fabric inspection method, system and electronic device. By identifying the barcode of the fabric roll to be inspected, the electronic information of the fabric defects is recorded by the smart device through algorithm calculation according to the rules. This scheme not only generates the fabric inspection report quickly and efficiently, but also can be connected with the warehouse and business system of the factory in real time, so that the production problems can be quickly found and solved.

[0065] The smart fabric inspection method, system and electronic device disclosed in the present application will be further described in detail below with reference to the accompanying drawings.

[0066] In combination with Figure 1 The smart fabric inspection method disclosed in the embodiment comprises the following steps:

[0067] In step S102, the barcode of the fabric roll to be inspected on the fabric inspection machine is scanned, and the fabric roll specification is determined, so as to display the defect recording template corresponding to the fabric specification on the smart device. The defect recording template is matched with the corresponding fabric roll specification, so as to facilitate data statistics and score calculation.

[0068] In step S104, a plurality of image pictures of part of the fabric spread on the fabric inspection area are sequentially obtained under different light and shadow environments. Defects will show different effects under different light and shadow conditions, and some defects do not show under white light conditions. Therefore, all defects can be found under different light and shadow conditions.

[0069] Step S106, a fabric defect detection model based on deep learning is used to identify fabric defects in each image picture in turn, and the fabric defect set corresponding to the partial fabric is recorded in the corresponding template area of the defect input template one by one, and the fabric defect score is automatically judged according to the defect scoring standard; wherein the fabric defect detection model based on deep learning is a convolutional neural network, which uses the Faster R-CNN algorithm of semi-supervised learning, and the Faster R-CNN algorithm realizes fast and accurate defect detection by introducing the RPN network and the end-to-end training strategy. By identifying fabric defects through the model, the error caused by manual visual identification of fabric defects and subsequent recording in the defect input template is avoided, and automatic score management is also realized.

[0070] Step S108, receiving a defect input end signal, generating a fabric score report according to the fabric evaluation standard, each fabric defect set and the score of all fabric defects automatically judged according to the defect scoring standard; wherein the fabric score report lists the total score of the fabric roll, the number of different fabric defects and the total score of each classified fabric defect; the scores of different fabric defects are different, for example, the score of a hole is higher than that of a black dot.

[0071] Step S110, according to the fabric grade evaluation standard and the fabric score report, the fabric roll is classified according to the fabric grade evaluation standard and the fabric score report. The purpose of fabric roll classification is to manage the inventory on one hand, and to help business development on the other hand, such as accepting orders according to the existing inventory or planning production tasks according to the inventory.

[0072] The intelligent fabric inspection method disclosed in the embodiment of the application can quickly generate a fabric inspection report by inputting fabric defects into the system through a traditional fabric inspection machine combined with intelligent equipment such as a handheld PAD terminal, which not only improves the input efficiency, but also saves the time cost of manual re-input into the system.

[0073] To improve the accuracy of fabric score in the fabric inspection process, the intelligent fabric inspection method disclosed in the embodiment also includes the position of the accurate fabric defect, and then deducts the score according to the corresponding score of the different position of the defect; the specific implementation process is as follows Figure 2As shown, the method comprises: step S202, setting a positioning template matching the inspection area according to the specifications of the fabric roll and the size of the inspection area of the cloth inspection machine, the positioning template comprising a plurality of positioning areas; step S204, for each fabric roll specification, configuring a defect entry template corresponding to the positioning template corresponding to the corresponding fabric roll specification on the intelligent device, the defect entry template setting a plurality of defect positioning areas; the existence of the defect positioning area can on one hand accurately identify the position of the fabric defect so as to process the fabric subsequently, and on the other hand the defect scores of different positions are different; step S206, establishing a positioning coordinate for the image picture so that the first position information of the fabric defect is located when the cloth defect detection model based on deep learning is identified and acquired, and the second position information of the fabric defect on the positioning template is determined on the cloth inspection machine; wherein the first position information is the coordinate information in the positioning coordinate where the fabric defect is located, and the second position information is the positioning area where the fabric defect is located; step S208, judging the position deviation of the first position information and the second position information, and when the position deviation is within a preset range, selecting the corresponding defect entry template according to the specifications of the fabric roll, the first position information or the second position information, and entering the fabric defect at the corresponding position of the defect entry template. By judging the position deviation of the first position information and the second position information, on one hand the detection accuracy of the cloth defect detection model based on deep learning can be verified, and on the other hand the data information of the positioning template and the image picture can be matched.

[0074] The process of accurately calculating the deduction score in the defect entry template further comprises: step S302, dividing a plurality of scoring zones according to the defect scoring standard, the scoring zones comprising a plurality of defect positioning areas; step S304, respectively adding corresponding deduction score items to each of the scoring zones according to the deduction scores of each fabric defect, the deduction score items comprising deduction scores corresponding to different fabric defects; wherein the deduction score items of each defect positioning area in each of the scoring zones are the same, as shown in detail in Figure 3 The deduction scores of the fabric defects in the same scoring zone are the same, which facilitates the work of the defect entry template; for example, the deduction score of the fabric defect in the middle of the fabric is higher than that of the fabric defect at the edge of the fabric.

[0075] Optionally, in order to facilitate the intelligent device and the cloth inspection machine to synchronize the identification and entry of the fabric defect, the defect entry template is designed as a format of a flip text in the embodiment, and the working process is designed as Figure 4As shown, the method comprises: step S402, determining the length of the fabric wound on the fabric roll according to the fabric roll specification; step S404, dividing the defect entry template corresponding to the fabric roll specification into a page-turning text according to the length of the fabric, any page of the page-turning text is similar to the fabric inspection area size, and each page of the page-turning text corresponds to each fabric segment photographed by the fabric roll; and step S406, when the fabric in the fabric inspection area of the fabric inspection machine is detected, sequentially releasing the fabric of the fabric roll to re-cover the fabric inspection area, and automatically switching to the next page after the corresponding defect entry template completes defect entry of the current page. The display page of the smart device does not need to be manually switched. In the embodiment, the smart device is designed as a handheld PAD terminal with a scanning function, and the handheld PAD terminal automatically jumps to the entry interface of the corresponding defect entry template after scanning the barcode of the fabric roll.

[0076] Optionally, the smart fabric inspection method further comprises performing qualified judgment on the fabric roll in the fabric inspection, and performing grade division on the qualified fabric roll after the fabric inspection is completed; specifically as Figure 5 As shown, the method comprises: step S502, when the fabric inspection machine is working, judging whether the fabric roll in the current inspection belongs to a qualified fabric roll in real time, and when the fabric roll belongs to a qualified fabric roll, performing grade division on the fabric roll according to the fabric grade evaluation standard and the fabric score report; wherein the process of judging whether the fabric roll in the current inspection belongs to a qualified fabric roll in real time comprises: calculating the fabric score of the fabric roll currently inspected by the fabric inspection machine in real time, and when the fabric score of the fabric roll in each preset inspection period is lower than the corresponding score threshold, the fabric roll is unqualified; this process is mainly aimed at poor quality fabric, and products are unqualified due to too many fabric defects, so such products are directly classified during the fabric inspection process, which not only can track the same batch of products for problem tracing, but also can reduce the calculation and classification after the fabric inspection. The types, quantities and corresponding areas of the entered fabric defects are calculated in real time, and when the number of target fabric defects exceeds the preset target or the area exceeds the preset proportion, the fabric roll is unqualified; this process aims to important fabric defects that affect the quality of the fabric, and when the important fabric defects appear or the area exceeds the preset proportion, such unqualified fabric roll is directly managed in the warehouse or problem tracing is performed.

[0077] In the embodiment of the present application, an electronic device is also provided, which comprises a computer program stored in a computer readable storage medium; when a processor of the electronic device reads the computer program from the computer readable storage medium, the processor executes the computer program, so that the electronic device executes the steps of the smart fabric inspection method described above. Taking the electronic device running on a computer as an example, as Figure 7As shown, the electronic device can include one or more (only one is shown in the figure) processors (the processor can include but not limited to a microprocessor MCU or a programmable logic device FPGA processing device), a memory for storing data, and a transmission device for communication function. Those skilled in the art can understand that, Figure 7 The structure shown is only schematic, which does not limit the structure of the above-mentioned electronic device.

[0078] The above-mentioned program can run in the processor, or also can be stored in the memory, that is, the computer readable medium, the computer readable medium includes permanent and non-permanent, removable and non-removable media can be realized by any method or technology to store information. Information can be computer readable instructions, data structure, program module or other data. Examples of computer storage media include, but are not limited to, phase change memory (PRAM), static random access memory (SRAM), dynamic random access memory (DRAM), other types of random access memory (RAM), read only memory (ROM), electrically erasable programmable read only memory (EEPROM), flash memory or other memory technology, compact disc read only memory (CD-ROM), digital versatile disc (DVD) or other optical storage, magnetic cassette tape, magnetic tape magnetic disk storage or other magnetic storage device or any other non-transmission medium, which can be used to store information that can be accessed by a computing device. According to the definition in this paper, computer readable medium does not include temporary computer readable medium, such as modulated data signal and carrier wave. These computer programs can also be loaded into a computer or other programmable data processing device, so that a series of operation steps are executed on the computer or other programmable device to generate computer implemented processing, so that the instructions executed on the computer or other programmable device provide steps for implementing the processes Figure 1 One flow or multiple flows and / or blocks Figure 1 The steps of the function specified in one block or multiple blocks correspond to different method steps, which can be realized by different modules.

[0079] In this embodiment, a device or system is provided, which can be called an intelligent cloth inspection system, such as Figure 6As shown, comprising: a scanning determination module for scanning the bar code of the fabric roll to be inspected on the cloth inspection machine, determining the fabric roll specification, so as to display the defect entry template corresponding to the fabric specification on the intelligent device; an identification acquisition module for sequentially acquiring multiple image pictures of part of the fabric spread on the cloth inspection area under different light environments; an entry evaluation module for adopting a cloth defect detection model based on deep learning to identify the fabric defects in each image picture in turn, merging the fabric defect sets corresponding to the part of the fabric, and entering the fabric defects into the corresponding template area of the defect entry template one by one, and automatically judging the fabric defect deduction score according to the defect scoring standard; wherein the cloth defect detection model based on deep learning adopts a Faster R-CNN algorithm of semi-supervised learning; a receiving generation module for receiving a defect entry end signal, generating a fabric score report according to the fabric evaluation standard, each fabric defect set and the deduction score of all fabric defects automatically judged according to the defect scoring standard; wherein the fabric score report lists the total deduction score of the fabric roll, the number of different fabric defects and the total deduction score corresponding to each classified fabric defect; a classification module for classifying the fabric roll completed by the cloth inspection according to the fabric grade evaluation standard and the fabric score report.

[0080] The system is used to realize the steps of the intelligent cloth inspection method disclosed in the above embodiments, which have been described and will not be repeated here.

[0081] For example, the intelligent cloth inspection system further comprises: a setting module configured to set a positioning template matching the cloth inspection area according to the specifications of the fabric roll and the size of the cloth inspection area of the cloth inspection machine, the positioning template comprising a plurality of positioning areas; a configuration module configured to configure, for each fabric roll specification, a defect entry template corresponding to the positioning template corresponding to the corresponding fabric roll specification on the intelligent device, the defect entry template comprising a plurality of defect positioning areas; an acquisition and determination module configured to establish a positioning coordinate for the image picture, so that the first position information of the fabric defect is located when the fabric defect is identified and acquired by the cloth defect detection model based on deep learning, and the second position information of the fabric defect on the positioning template is determined on the cloth inspection machine; wherein the first position information is coordinate information in the positioning coordinate where the fabric defect is located, and the second position information is the positioning area where the fabric defect is located; a judgment and entry module configured to judge the positional deviation of the first position information and the second position information, and when the positional deviation is within a preset range, select the corresponding defect entry template according to the specifications of the fabric roll and the first position information or the second position information, and enter the fabric defect at the corresponding position of the defect entry template. Optionally, the embodiment designs the format of the defect entry template according to the specifications of the fabric roll to be inspected, for example, the format of the defect entry template is a page-turning text, that is: according to the specifications of the fabric roll, the length of the fabric wound thereon is determined; according to the length of the fabric, the defect entry template corresponding to the fabric roll specification is divided into a page-turning text, any page of the page-turning text is similar in size to the cloth inspection area; after the cloth inspection area of the cloth inspection machine completes the image picture acquisition and performs defect detection and identification, the fabric of the fabric roll is sequentially released to re-cover the cloth inspection area, and the corresponding defect entry template automatically switches to the next page after completing the defect entry of the current page.

[0082] For another example, the intelligent cloth inspection system further comprises: a division unit configured to divide the defect entry template into a plurality of scoring zones according to a defect scoring standard, the scoring zones comprising a plurality of defect positioning areas; and an additional unit configured to respectively add corresponding deduction score items to each of the scoring zones according to the deduction scores of each fabric defect, the deduction score items comprising deduction scores of different fabric defects; wherein the deduction score items of each defect positioning area in each scoring zone are the same.

[0083] For example, the intelligent cloth inspection system further comprises a judgment module configured to determine whether the fabric roll in current inspection is a qualified fabric roll when the cloth inspection machine is working, and to divide the fabric roll into different grades according to the fabric grade evaluation standard and the fabric score report when the fabric roll is a qualified fabric roll. The process of determining whether the fabric roll in current inspection is a qualified fabric roll comprises: calculating the fabric score of the fabric roll in current inspection in real time, and determining that the fabric roll is unqualified when the fabric score of the fabric roll in each preset inspection period is lower than the corresponding score threshold. The process of determining whether the fabric roll is unqualified further comprises: calculating the type, quantity and corresponding area of the input fabric defects in real time, and determining that the fabric roll is unqualified when the quantity of the target fabric defect exceeds the preset target or the area exceeds the preset proportion. The process of determining whether the fabric roll is unqualified in the cloth inspection process is beneficial to finding the production or storage defects that cause product quality problems and to repairing in advance.

[0084] The intelligent cloth inspection method, system and electronic device disclosed in the application combine the cloth inspection machine and the intelligent device, solve the problems of low efficiency of manually recording fabric defects and errors in secondary input during cloth inspection, realize rapid positioning of fabric defects on fabric rolls of different specifications and determination of the deduction score through the design of the positioning template and the defect input template, generate the score report through the deduction score item attached to each defect positioning area, and determine whether the fabric roll in current inspection is qualified and the grade of the fabric roll. In addition, the intelligent cloth inspection method can be used for inventory management, and the collection of information of all fabric rolls in the inventory is helpful for business planning of the manufacturer.

[0085] Although the application has been disclosed as above with reference to the preferred embodiments, the application is not limited thereto. Any modification and improvement made by those skilled in the art without departing from the spirit and scope of the application shall fall within the protection scope of the application. The protection scope of the application shall be defined by the claims.

Claims

1. A method for intelligent fabric inspection, characterized in that, include: Scan the barcode of the fabric roll to be inspected on the fabric inspection machine to determine the fabric roll specifications so that the corresponding fabric specification defect entry template can be displayed on the smart device. Under different lighting conditions, multiple images of the fabric roll laid out on the upper part of the fabric inspection area were acquired sequentially. A deep learning-based fabric defect detection model is used to sequentially identify fabric defects in each image. After merging them into a set of fabric defects for the corresponding part of the fabric, the fabric defects are entered one by one in the corresponding template area of ​​the defect entry template. The fabric defects are automatically judged and deducted points according to the defect scoring standard. The deep learning-based fabric defect detection model adopts the semi-supervised learning Faster R-CNN algorithm. Upon receiving the defect entry completion signal, the system generates a fabric score report based on the fabric evaluation criteria, the defect sets of each fabric, and the deduction scores for all fabric defects automatically judged according to the defect scoring criteria. The fabric score report lists the total deduction score of the fabric roll, the number of defects in different fabrics, and the total deduction score corresponding to the defects in each category of fabrics. Based on the fabric grade evaluation standards and fabric score reports, the fabric rolls that have completed the fabric inspection are classified into grades. The intelligent fabric inspection method further includes: According to the specifications of the fabric roll and the size of the fabric inspection area of ​​the fabric inspection machine, a positioning template matching the fabric inspection area is set, and the positioning template includes several positioning areas. For each fabric roll specification, a defect entry template corresponding to the positioning template of the corresponding fabric roll specification is configured on the intelligent device, and the defect entry template is set with several defect positioning areas. The image is established with positioning coordinates so that when the deep learning-based fabric defect detection model identifies and acquires fabric defects, it can simultaneously locate the first position information of the fabric defect and determine its second position information on the positioning template on the fabric inspection machine; wherein, the first position information is the coordinate information in the positioning coordinates of the fabric defect, and the second position information is the positioning area of ​​the fabric defect. Determine the positional deviation of the first position information and the second position information. When the positional deviation is within a preset range, select the corresponding defect entry template according to the specifications of the fabric roll, the first position information or the second position information, and enter the fabric defect at the corresponding position of the defect entry template.

2. The intelligent fabric inspection method according to claim 1, characterized in that, Also includes: According to the defect scoring criteria, the defect entry template is divided into several scoring areas, and the scoring areas include several defect location areas. Based on the deduction points for each fabric defect, corresponding deduction points are added to each of the scoring areas. The deduction points include deduction points for different fabric defects. The deduction points for each defect location area within each scoring area are the same.

3. The intelligent fabric inspection method according to claim 1, characterized in that, Also includes: Determine the length of the fabric to be rolled up based on the fabric roll specifications. Based on the length of the fabric, the defect entry template corresponding to the specifications of the fabric roll is divided into page-turning text. Each page of the page-turning text is similar in size to the fabric inspection area, and each page of the page-turning text corresponds to each fabric segment photographed from the fabric roll. After the fabric in the fabric inspection area of ​​the fabric inspection machine has been inspected, the fabric rolls are released sequentially to cover the inspection area again, and the corresponding defect entry template automatically switches to the next page after completing the defect entry for the current page.

4. The intelligent fabric inspection method according to claim 1, characterized in that, Also includes: When the fabric inspection machine is working, it determines in real time whether the fabric roll being inspected is a qualified fabric roll, and when the fabric roll is a qualified fabric roll, it classifies the fabric roll into grades according to the fabric grade evaluation standard and fabric score report. The process of determining in real time whether the fabric roll being inspected is a qualified fabric roll includes: The fabric score of the fabric roll being inspected by the fabric inspection machine is calculated in real time. If the fabric score of the inspected fabric roll is lower than the corresponding score threshold in each preset inspection cycle, the fabric roll is unqualified. The type, quantity, and corresponding area of ​​fabric defects are calculated and recorded in real time. When the number of defects in the target fabric exceeds the preset target or the area exceeds the preset ratio, the fabric roll is deemed unqualified.

5. A smart fabric inspection system, characterized in that, include: The scanning and determination module is used to scan the barcode of the fabric roll to be inspected on the fabric inspection machine to determine the specifications of the fabric roll so that the corresponding defect entry template can be displayed on the smart device. The identification and acquisition module is used to sequentially acquire multiple image images of the fabric roll spread out on the upper part of the fabric inspection area under different lighting and shadow environments. The input and evaluation module is used to sequentially identify fabric defects in each image using a deep learning-based fabric defect detection model, merge them into a set of fabric defects for the corresponding part of the fabric, and then input the fabric defects one by one into the corresponding template area of ​​the defect input template. The module also automatically evaluates the fabric defects and deducts points according to the defect scoring criteria. The deep learning-based fabric defect detection model uses the semi-supervised learning Faster R-CNN algorithm. The receiving and generating module is used to receive the defect entry end signal, and generate a fabric score report based on the fabric evaluation standard, the set of defects of each fabric, and the deduction points of all fabric defects automatically judged according to the defect scoring standard; wherein, the fabric score report lists the total deduction points of the fabric roll, the number of defects of different fabrics, and the total deduction points corresponding to the defects of each category of fabrics. The classification module is used to classify the fabric rolls that have completed the fabric inspection according to the fabric grade evaluation standards and fabric score reports. The intelligent fabric inspection system also includes: The setting module is used to set a positioning template that matches the fabric inspection area according to the specifications of the fabric roll and the size of the fabric inspection area of ​​the fabric inspection machine. The positioning template includes several positioning areas. The configuration module is used to configure a defect entry template corresponding to the positioning template of the corresponding fabric roll specification on the smart device for each fabric roll specification. The defect entry template is set with several defect positioning areas. The acquisition and determination module is used to establish positioning coordinates for the image so that when the deep learning-based fabric defect detection model identifies and acquires fabric defects, it can simultaneously locate the first position information of the fabric defect and determine its second position information on the positioning template on the fabric inspection machine; wherein, the first position information is the coordinate information in the positioning coordinates of the fabric defect, and the second position information is the positioning area where the fabric defect is located. The judgment and input module is used to judge the position deviation of the first position information and the second position information. When the position deviation is within a preset range, the module selects the corresponding defect input template according to the specifications of the fabric roll, the first position information or the second position information, and inputs the fabric defect at the corresponding position of the defect input template.

6. The intelligent fabric inspection system according to claim 5, characterized in that, Also includes: A division unit is used to divide the defect entry template into several scoring areas according to the defect scoring criteria, wherein the scoring area includes several defect location areas; An additional unit is used to add corresponding deduction items to each of the scoring areas according to the deduction points for each fabric defect. The deduction items include deduction points for different fabric defects. The deduction items for each defect location area in each of the scoring areas are the same.

7. The intelligent fabric inspection system according to claim 5, characterized in that, Also includes: The judgment module is used to determine in real time whether the fabric roll being inspected is a qualified fabric roll when the fabric inspection machine is working, and to classify the fabric roll into grades according to the fabric grade evaluation standard and fabric score report when the fabric roll is a qualified fabric roll. The process of determining in real time whether the fabric roll being inspected is a qualified fabric roll includes: The fabric score of the fabric roll being inspected by the fabric inspection machine is calculated in real time. If the fabric score of the inspected fabric roll is lower than the corresponding score threshold in each preset inspection cycle, the fabric roll is unqualified. The type, quantity, and corresponding area of ​​fabric defects are calculated and recorded in real time. When the number of defects in the target fabric exceeds the preset target or the area exceeds the preset ratio, the fabric roll is deemed unqualified.

8. An electronic device, characterized in that, The method includes a computer program stored in a computer-readable storage medium; when a processor of an electronic device reads the computer program from the computer-readable storage medium, the processor executes the computer program, causing the electronic device to perform the steps of the method for intelligent fabric verification according to any one of claims 1-4.

Images