Fabric inspection processing method, device, equipment, system and medium

Abstract

The present invention provides a cloth inspection processing method, device, equipment, system and medium. By inputting the cloth image sent by the cloth inspection machine into the cloth detection model, the defect type and position in the cloth image are identified, marked in the cloth image and then sent to the user's terminal device. After the user determines that the position and type of the marked defect are correct, a stop cloth inspection instruction is sent to the server, and then the server sends a stop cloth inspection instruction to the cloth inspection machine, and the cloth inspection machine can stop the rotating shaft. After that, the user sends a marking message to the server, and then the server processes the cloth inspection message to obtain the defect coordinates and sends a marking instruction to the cloth inspection machine, and the cloth inspection machine marks the cloth according to the defect coordinates. This solution realizes the detection of cloth through the combination of manual operation and the cloth detection model, effectively reducing the cloth inspection cost and improving the cloth inspection efficiency.

Description

Technical Field

[0001] The present invention relates to the field of cloth inspection, and particularly to a cloth inspection processing method, device, equipment, system and medium. Background Art

[0002] In daily life, people's clothes, bed sheets, backpacks, etc. are all inseparable from cloth, and quality inspections are also carried out when purchasing clothes, bed sheets, backpacks, etc. Therefore, cloth manufacturers will inspect the cloth after it is produced.

[0003] In the prior art, manufacturers will train new employees through manual instruction. After the training is completed, the cloth inspection user will detect the appearance defects of the cloth on the cloth inspection machine. The cloth is placed on the cloth inspection machine and rolled for display. After the cloth inspection user observes the defect, they will mark it to complete the inspection. In addition, when the order volume is large, the existing number of cloth inspection users in the manufacturer may not meet the cloth inspection requirements.

[0004] In summary, the existing cloth inspection method relies entirely on manual training and inspection, resulting in high cloth inspection costs and low efficiency. Summary of the Invention

[0005] The present invention provides a cloth inspection processing method, device, equipment, system and medium, which are used to solve the problems of high cloth inspection costs and low efficiency in the prior art.

[0006] In a first aspect, the present invention provides a cloth inspection processing method, which is applied to a server, and the method includes:

[0007] During the cloth inspection process, receive the cloth image to be processed sent by the cloth inspection machine;

[0008] Input the cloth image into a pre-acquired cloth detection model to determine the defect position and defect type in the cloth image. The cloth detection model is used to identify the position and type of cloth defects in the cloth image;

[0009] Mark in the cloth image according to the defect position and defect type in the cloth image to obtain a marked cloth image;

[0010] Send the marked cloth image to the user's terminal device;

[0011] After receiving the stop cloth inspection instruction sent by the terminal device, send the stop cloth inspection instruction to the cloth inspection machine. The stop cloth inspection instruction is an instruction generated by the terminal device in response to the user's operation for controlling the cloth inspection machine to stop rotating the shaft.

[0012] In a specific embodiment, before receiving the to-be-processed cloth image sent by the cloth inspection machine during the cloth inspection process, the method further includes:

[0013] Receiving the start cloth inspection instruction sent by the terminal device, and sending the start cloth inspection instruction to the cloth inspection machine, where the start cloth inspection instruction is used to control the cloth inspection machine to start the rotating shaft and start cloth inspection.

[0014] In a specific embodiment, the method further includes:

[0015] Obtaining a pre-calibrated set of cloth images, where the set of cloth images includes calibrated cloth images of multiple marked cloths with defect positions and defect types;

[0016] According to the set of cloth images, using a convolutional network model for model training to obtain the cloth detection model.

[0017] In a specific embodiment, the method further includes:

[0018] Calculating the coordinate information of the at least one defect according to the defect marking message;

[0019] According to the coordinate information of the at least one defect, sending a marking instruction to the cloth inspection machine, where the marking instruction includes the coordinate information of the at least one defect, and the marking instruction is used to control the cloth inspection machine to mark the at least one defect on the cloth.

[0020] In a specific embodiment, the method further includes:

[0021] Receiving the cloth inspection requirement information sent by the factory cloth inspection platform, where the cloth inspection requirement information includes the rating of the required cloth inspection users and the number of cloth inspection users;

[0022] According to the pre-obtained rating information of each user and the cloth inspection requirement information, obtaining at least one target cloth inspection user that meets the cloth inspection requirement information from the currently selectable cloth inspection users;

[0023] Sending a cloth inspection task to the terminal devices corresponding to the at least one target cloth inspection user.

[0024] In a specific embodiment, the rating information of each cloth inspection user includes the rating of the cloth inspection user, and the video playback speed, false negative rate, and false positive rate corresponding to the rating; the rating information is obtained according to the cloth inspection data of the cloth inspection user during the cloth inspection training process.

[0025] In a second aspect, the present invention provides a cloth inspection processing method applied to a terminal device, and the method includes:

[0026] During the fabric inspection process, receive the marked fabric image sent by the server, where the marked fabric image includes the defect positions and defect types automatically detected and marked by the server;

[0027] Display the marked fabric image on the graphical user interface;

[0028] In response to the user's operation on the stop fabric inspection button on the graphical user interface, send a stop fabric inspection instruction to the server, where the stop fabric inspection instruction is used to control the fabric inspection machine to stop the rotating shaft.

[0029] In a specific implementation manner, before receiving the marked fabric image sent by the server during the fabric inspection process, the method further includes:

[0030] In response to the user's operation on the start fabric inspection button on the graphical user interface, send a start fabric inspection instruction to the server, where the start fabric inspection instruction is used to control the fabric inspection machine to start the rotating shaft and start fabric inspection.

[0031] In a specific implementation manner, the method further includes:

[0032] In response to the positions and types of at least one defect manually marked by the user on the graphical user interface, send defect marking information to the server, where the defect marking information includes the positions and types of the at least one defect.

[0033] In a specific implementation manner, the method further includes:

[0034] Receive the fabric inspection task sent by the server and display the fabric inspection task on the graphical user interface.

[0035] In a specific implementation manner, the method further includes:

[0036] During the fabric inspection training process, in response to the user's operation on the graphical user interface, obtain fabric inspection data;

[0037] Send the fabric inspection data to the server, where the fabric inspection data is used to obtain the user's rating information, and the rating information includes the user's rating, as well as the video playback speed, missed detection rate, and false alarm rate corresponding to the rating.

[0038] In a third aspect, the present invention provides a fabric inspection processing device, including:

[0039] A receiving module, configured to receive the fabric image to be processed sent by the fabric inspection machine during the fabric inspection process;

[0040] A detection module, configured to input the cloth image into a pre-acquired cloth detection model to determine the defect positions and defect types in the cloth image, where the cloth detection model is used to identify the positions and types of cloth defects in the cloth image;

[0041] A processing module, configured to mark the cloth image according to the defect positions and defect types in the cloth image to obtain a marked cloth image;

[0042] A sending module, configured to send the marked cloth image to the user's terminal device;

[0043] The sending module is further configured to, after receiving the stop cloth inspection instruction sent by the terminal device, send the stop cloth inspection instruction to the cloth inspection machine, where the stop cloth inspection instruction is an instruction generated by the terminal device in response to the user's operation and used to control the cloth inspection machine to stop rotating its shaft.

[0044] In a fourth aspect, the present invention provides a cloth inspection processing device, including:

[0045] A receiving module, configured to receive the marked cloth image sent by the server during the cloth inspection process, where the marked cloth image includes the defect positions and defect types automatically detected and marked by the server;

[0046] A display module, configured to display the marked cloth image on a graphical user interface;

[0047] A sending module, configured to, in response to the user's operation on the stop cloth inspection button on the graphical user interface, send a stop cloth inspection instruction to the server, where the stop cloth inspection instruction is an instruction used to control the cloth inspection machine to stop rotating its shaft.

[0048] In a fifth aspect, the present invention provides a server, including:

[0049] A processor, a memory, and a communication interface;

[0050] The memory is used to store executable instructions of the processor;

[0051] Wherein, the processor is configured to execute the cloth inspection processing method according to any one of the first aspect by executing the executable instructions.

[0052] In a sixth aspect, the present invention provides a terminal device, including:

[0053] A processor, a memory, a display, and a communication interface;

[0054] The memory is used to store executable instructions of the processor;

[0055] Wherein, the processor is configured to execute the fabric inspection processing method according to any one of the second aspect by executing the executable instructions.

[0056] In a seventh aspect, the present invention provides a fabric inspection processing system, including:

[0057] A server, a terminal device, a switch, and a fabric inspection machine;

[0058] The server is used to execute the fabric inspection processing method according to any one of the first aspect;

[0059] The terminal device is used to execute the fabric inspection processing method according to any one of the second aspect;

[0060] The switch is used to transmit network communication signals;

[0061] The fabric inspection machine is used to receive the start fabric inspection instruction, the stop fabric inspection instruction, and the marking instruction sent by the server, control the rotation shaft on the fabric inspection machine to start rotating according to the start fabric inspection instruction, control the rotation shaft on the fabric inspection machine to stop rotating according to the stop fabric inspection instruction, and control the coding machine on the fabric inspection machine to mark the defects on the fabric according to the marking instruction;

[0062] The fabric inspection machine is further used to send the image of the fabric to be processed to the server.

[0063] In an eighth aspect, the present invention provides a readable storage medium, on which a computer program is stored, and when the computer program is executed by a processor, it implements the fabric inspection processing method according to any one of the first to second aspects.

[0064] The fabric inspection processing method, device, equipment, system and medium provided by the present invention input the fabric image sent by the fabric inspection machine into the fabric detection model, identify the type and position of the defects in the fabric image, mark them in the fabric image and then send them to the user's terminal device. After the user determines that the position and type of the marked defects are correct, the user sends an instruction to stop fabric inspection to the server, and then the server sends a stop fabric inspection instruction to the fabric inspection machine, and the fabric inspection machine can stop the rotation shaft. Then the user sends a marking message to the server, and then the server processes the fabric inspection message to obtain the defect coordinates and sends a marking instruction to the fabric inspection machine, and the fabric inspection machine marks the fabric according to the defect coordinates. This solution realizes the detection of the fabric through the combination of manual work and the fabric detection model, effectively reducing the fabric inspection cost and improving the fabric inspection efficiency. Description of the Drawings

[0065] To more clearly illustrate the technical solutions in the embodiments of the present invention or the prior art, the following will briefly introduce the drawings required for the description of the embodiments or the prior art. Obviously, the drawings in the following description are some embodiments of the present invention. For those of ordinary skill in the art, without creative efforts, other drawings can also be obtained based on these drawings.

[0066] Figure 1 Schematic diagram of the application scenario of the cloth inspection processing method provided by the present invention;

[0067] Figure 2a Schematic diagram of the process of Embodiment 1 of the cloth inspection processing method provided by the present invention;

[0068] Figure 2b Physical diagram of the encoder provided by the present invention;

[0069] Figure 2c Defect type diagram provided by the present invention;

[0070] Figure 2d Marked cloth image provided by the present invention;

[0071] Figure 3a Schematic diagram of the process of Embodiment 2 of the cloth inspection processing method provided by the present invention;

[0072] Figure 3b Schematic diagram of a user marking a defect provided by the present invention;

[0073] Figure 3c Schematic diagram of calculating a scale factor provided by the present invention;

[0074] Figure 3d Schematic diagram of the relationship between two cloth inspection methods provided by the present invention;

[0075] Figure 4a Schematic diagram of the process of Embodiment 3 of the cloth inspection processing method provided by the present invention;

[0076] Figure 4b Schematic diagram of a pre-calibrated cloth image provided by the present invention;

[0077] Figure 5 Schematic diagram of the structure of Embodiment 1 of the cloth inspection processing device provided by the present invention;

[0078] Figure 6a Schematic diagram of the structure of Embodiment 2 of the cloth inspection processing device provided by the present invention;

[0079] Figure 6b Schematic diagram of the structure of Embodiment 3 of the cloth inspection processing device provided by the present invention;

[0080] Figure 7A schematic diagram of the structure of a server provided by the present invention;

[0081] Figure 8 A schematic diagram of the structure of a terminal device provided by the present invention. DETAILED DESCRIPTION

[0082] In order to make the purpose, technical solution and advantages of the embodiment of the present invention clearer, the technical solution in the embodiment of the present invention will be clearly and completely described below in conjunction with the drawings in the embodiment of the present invention. Obviously, the described embodiment is a part of the embodiment of the present invention, not all the embodiments. Based on the embodiment of the present invention, all other embodiments made by ordinary technicians in this field under the enlightenment of this embodiment belong to the scope of protection of the present invention.

[0083] The terms "first", "second", "third", "fourth", etc. (if any) in the specification and claims of the present invention and the above-mentioned drawings are used to distinguish similar objects, and are not necessarily used to describe a specific order or sequence. It should be understood that the data used in this way can be interchanged where appropriate, so that the embodiments of the present invention described herein can be implemented in an order other than those illustrated or described herein. In addition, the terms "including" and "having" and any variations thereof are intended to cover non-exclusive inclusions, for example, a process, method, system, product or device that includes a series of steps or units is not necessarily limited to those steps or units that are clearly listed, but may include other steps or units that are not clearly listed or inherent to these processes, methods, products or devices.

[0084] In daily life, cloth is widely used. The clothes people wear and the bed sheets they use are all inseparable from cloth. There are also strict requirements on the quality of cloth. Therefore, manufacturers will test the cloth after production is completed.

[0085] Since in the current cloth quality inspection industry, manual inspection operations are generally carried out on cloth inspection machines, the inspected width is at least more than 1.5 meters, and the cloth runs relatively fast, usually 20 - 50 meters per minute. Working on the cloth inspection machine for a long time, the human eye is prone to fatigue, resulting in missed and misjudged inspections of subtle and unobvious defects. In addition, the inspection standards for cloth are not uniform, and it entirely depends on people's subjective experience for learning and judgment. Currently, there are also intelligent cloth inspection machines in the industry, but facing the diversity of cloth and defects, the machine still cannot completely replace humans to complete the cloth inspection work. In addition, current cloth inspection users are generally recruited and trained by factories themselves. As a result, the short-term output is easily affected by order changes, personnel flow, etc., and it is impossible to maximize production and value. Before cloth inspection users take up their posts, the training completely relies on experienced old masters to orally impart quality inspection experience, which cannot achieve the standardization of defects, resulting in differences in the quality inspection results between people, and the training time and cost are relatively high. Therefore, the existing technology leads to relatively high cloth inspection costs and low cloth inspection efficiency.

[0086] In the process of researching the relevant system for cloth inspection, the inventor found that the cloth image can be collected by a camera, the cloth image can be input into a cloth detection model, and after obtaining the defect type and position, it can be displayed to the user. The user checks and marks the defect type and position obtained by the cloth detection model, and can also use the defect data marked manually to strengthen the training of the cloth detection model, so that the recognition accuracy of the cloth detection model is higher. In addition, the cloth inspection video can be collected in advance. The personnel to be trained select the playback speed to watch the video and mark the defect positions. According to the playback speed, false negative rate and false positive rate, the personnel to be trained can be rated to achieve the training of employees. After the training is completed, the cloth inspection users can also be scheduled according to the requirements of the manufacturer, without the factory spending time and cost to maintain the recruitment and training tasks of quality inspection personnel. The manpower allocation is relatively flexible. When the production demand is large, the shortage of manpower in the factory can be quickly filled. When the product quality requirements are high, multiple people can carry out cloth inspection operations simultaneously, so as to complete the cloth inspection task on time and with quality assurance. Based on the above inventive concept, the cloth inspection processing solution in the present invention is designed.

[0087] Before the cloth inspection operation, it is necessary to train and schedule the cloth inspection users. The following will explain the training and scheduling method for the cloth inspection users provided by the present invention.

[0088] The training of cloth inspection users specifically includes the following steps:

[0089] During the cloth inspection training process, the terminal device responds to the user's operation on the graphical user interface and obtains cloth inspection data.

[0090] Before training the fabric inspection users, the staff need to input multiple fabric inspection videos and their corresponding accuracy rates collected into the server so that the fabric inspection users can obtain the videos for viewing. The fabric inspection videos contain standard samples of various fabrics and defects. The positions and types of defects on the screen are recorded in advance and used as learning prompts or reference answers. The fabric inspection videos can be obtained from the network or be videos recorded when using the fabric inspection method provided by the present invention. The present invention does not specifically limit the fabric inspection videos, which can be obtained according to the actual situation.

[0091] The server screens the fabric inspection videos input by the staff according to whether the fabric inspection accuracy rate reaches a preset value. After the screening is completed, the fabric inspection videos are numbered and stored. Exemplarily, the server screens out the videos with an accuracy rate greater than or equal to 95% for numbering and storage, which can ensure the accuracy of calculating the false negative rate and false positive rate of the fabric inspection users. The preset value is set by the staff. The present invention does not specifically limit the size of the preset value, which can be set according to the actual situation.

[0092] In this step, after storing the fabric inspection videos in the server, the fabric inspection users can be trained. First, the fabric inspection users need to input their account passwords on the terminal to log in to the system, and the terminal sends the account passwords to the server for verification. Logging in with the account password can ensure personal operation and also make the account correspond to the level of the fabric inspection users after training.

[0093] After the server verifies and passes, it sends an instruction allowing login to the terminal, and the fabric inspection users can further operate.

[0094] The fabric inspection users select the number of the fabric inspection video and the video playback speed on the user graphical interface of the terminal. The terminal sends the fabric inspection video number and the playback speed to the server. The server can find the corresponding fabric inspection video according to the number and send it to the terminal. The fabric inspection users can watch the fabric inspection video at the playback speed they choose on the terminal and mark the defects. Exemplarily, at the beginning, when the fabric inspection users are not familiar, they can choose the fabric inspection speed of 10 meters per minute, and the actual video playback speed is reduced to 0.5 times the speed. When the staff gradually get familiar, it can be adjusted to the normal speed of 20 meters per minute and the video at 1 times the speed. While the video is playing, the fabric inspection users directly mark on the screen. For example, when a defect is found, the screen is stopped, and the type and position of the defect are marked. In this way, the terminal device can obtain the fabric inspection data.

[0095] The terminal device sends the fabric inspection data to the server. The fabric inspection data is used to obtain the rating information of the user. The rating information includes the rating of the user, as well as the video playback speed, false negative rate, and false positive rate corresponding to the rating. The server obtains the rating information of the user according to the fabric inspection data.

[0096] In this step, after the fabric inspection user marks the type and location of the defect on the terminal, the terminal sends the fabric inspection data to the server. The fabric inspection data includes the type and location of the defect. Then, the server can calculate the omission rate and false alarm rate of this fabric inspection user, and rate the fabric inspection user according to the video playback speed, omission rate, and false alarm rate. After rating, a mapping relationship between the account and level of this fabric inspection user is established and stored, so as to find the fabric inspection user according to the level during personnel scheduling. Exemplarily, the correspondence between the video playback speed, omission rate, and false alarm rate and the level of the fabric inspection user can be shown in Table 1:

[0097] Table 1

[0098]

[0099]

[0100] It should be noted that Table 1 only shows the correspondence between the video playback speed, omission rate, and false alarm rate and the level of the fabric inspection user in the form of an example, and does not limit this correspondence. In actual application, this correspondence can be set according to actual needs.

[0101] It should be noted that the rating information of each fabric inspection user includes the rating of the fabric inspection user, as well as the video playback speed, omission rate, and false alarm rate corresponding to the rating; the rating information is obtained from the fabric inspection data of the fabric inspection user during the fabric inspection training process.

[0102] After the training of the fabric inspection user is completed, when a production enterprise needs a fabric inspection user to inspect the fabric, the fabric inspection user can be scheduled.

[0103] The scheduling of the fabric inspection user specifically includes the following steps:

[0104] The terminal device sends fabric inspection requirement information to the server through the factory fabric inspection platform on the terminal device. The server receives the fabric inspection requirement information sent by the factory fabric inspection platform. The fabric inspection requirement information includes the rating of the required fabric inspection user and the number of fabric inspection users.

[0105] In this step, when a production manufacturer needs to use the fabric inspection user trained by this solution, the fabric inspection user can be selected through the terminal, and the fabric inspection requirement information can be sent to the server through the factory fabric inspection platform. Then, the server can select the fabric inspection user according to the fabric inspection requirement information.

[0106] Exemplarily, when inspecting cloth, the images of the same piece of cloth can be sent to multiple cloth-inspecting users, and defect identification and inspection can be carried out independently online at the same time. The more people there are, the more guaranteed the quality of the cloth will be. Furthermore, the manufacturer can schedule personnel according to the orders of the factory. For example, the factory needs 10 cloth-inspecting users in January and 20 cloth-inspecting users in February according to the orders, and can also determine the number of different cloth-inspecting users, qualifications, false negative rates, and false positive rates according to the quality grades of the cloth. Exemplarily, the corresponding relationships among the cloth-inspecting users, false negative rates, and false positive rates and the cloth quality grades can be shown in Table 2 as follows:

[0107] Table 2

[0108] Fabric quality grade Fabric inspection user Omission rate False alarm rate Special grade 4 persons (senior) 1％ 1％ Fine quality 2 persons (above intermediate) 10％ 10％ Average 1 person (above junior) 20％ 20％

[0109] It should be noted that Table 1 only shows the corresponding relationships among the cloth-inspecting users, false negative rates, and false positive rates and the cloth quality grades in the form of examples, and does not limit such corresponding relationships. In actual application processes, such corresponding relationships can be set according to actual needs.

[0110] The server obtains at least one target cloth-inspecting user that meets the cloth-inspecting requirement information from the currently selectable cloth-inspecting users according to the pre-obtained rating information of each user and the cloth-inspecting requirement information.

[0111] In this step, after the server obtains the cloth-inspecting requirement information, it can select cloth-inspecting users according to the ratings and the number of cloth-inspecting users in the cloth-inspecting requirement information. Since the rating information of the cloth-inspecting users is stored in the server after the training process of the cloth-inspecting users is completed, the target cloth-inspecting users that meet the cloth-inspecting requirements can be found according to the rating information.

[0112] It should be noted that when there is only one piece of rating information stored in the server, the number of target cloth-inspecting users selected according to the cloth-inspecting requirements is only one. When there are multiple pieces of rating information stored in the server, the number of target cloth-inspecting users that can be selected according to the cloth-inspecting requirements is at least one.

[0113] The server sends a cloth-inspecting task to the terminal devices corresponding to at least one target cloth-inspecting user. The terminal devices receive the cloth-inspecting task sent by the server and display the cloth-inspecting task on the graphical user interface.

[0114] In this step, after the server selects the target cloth-inspecting users, it can send the cloth-inspecting task to the terminals corresponding to the cloth-inspecting users, so that the terminal devices display the cloth-inspecting task on the user graphical interface and present it to the cloth-inspecting users, and the cloth-inspecting users carry out cloth inspection according to the cloth-inspecting task.

[0115] After the training and scheduling of the fabric inspection users are completed, the manufacturer can let the fabric inspection users participate in fabric inspection. The fabric inspection processing method provided by the present invention will be described below.

[0116] Exemplarily, Figure 1 is a schematic diagram of the application scenario of the fabric inspection processing method provided by the present invention. As Figure 1 shown, the application scenario includes: a terminal device 11, a switch 12, a switch 13, a server 14, at least one industrial camera ( Figure 1 2 industrial cameras are shown, namely industrial camera 15 and industrial camera 16), an encoder 17, a coding machine 18, a fabric 19, a light source 20, a controller 21, and a rotating shaft 22. Among them, at least one industrial camera ( Figure 1 2 industrial cameras are shown, namely industrial camera 15 and industrial camera 16) and the light source 20 constitute an optical acquisition system, and the optical acquisition system, the encoder 17, the coding machine 18, the fabric 19, the controller 21, and the rotating shaft 22 constitute a fabric inspection machine.

[0117] Exemplarily, the terminal device 11 can send a start fabric inspection instruction, a stop fabric inspection instruction, and defect marking information to the server through the switch 12 and the switch 13, and can also receive the marked fabric image from the server through the switch 12 and the switch 13.

[0118] The switch 12 and the switch 13 can be connected through a network to transmit network communication signals, and the terminal device 12 and the server 14 can realize data transmission through the switch 12 and the switch 13.

[0119] The server 14 can receive the start fabric inspection instruction, the stop fabric inspection instruction, and the defect marking information from the terminal through the switch 12 and the switch 13, can send the marked fabric image through the switch 12 and the switch 13, can receive the to-be-processed fabric image sent by the industrial camera 15 in the fabric inspection machine, can receive the cloth length signal sent by the encoder 17 in the fabric inspection machine, can send a start fabric inspection instruction and a stop fabric inspection instruction to the controller 21, and can also send a marking instruction to the coding machine 18.

[0120] The industrial camera 15 can collect the to-be-processed fabric image and send the to-be-processed image to the server 14.

[0121] The encoder 17 can collect the cloth length information and send a cloth length signal to the server 14;

[0122] The coding machine 18 includes a one-dimensional moving platform and a label automatic pasting device, can receive the marking instruction from the server 14, and can paste a label on the fabric 19 according to the marking instruction.

[0123] The light source 20 can illuminate the cloth 19 to increase the brightness of the cloth 19.

[0124] The controller 21 contains a programmable logic controller (PLC) system for the cloth inspection machine, including the start and stop control of the rotating shaft of the cloth inspection machine. The controller 21 can receive the start cloth inspection instruction and the stop cloth inspection instruction from the server 14, and control the rotation and stop of the rotating shaft 22 according to the start cloth inspection instruction and the stop cloth inspection instruction.

[0125] The rotating shaft 22 can accept the control of the controller 21 and can also drive the cloth 19 to move and stop.

[0126] It should be noted that Figure 1 is only a schematic diagram of an application scenario provided by the embodiment of the present invention. The embodiment of the present invention does not Figure 1 limit the actual form and specific quantity of various devices included therein, nor does it Figure 1 limit the positional relationship and interaction mode between the devices included therein. In the specific application of the solution, it can be set according to actual needs.

[0127] It should be understood that the terminal device is a device that can run the corresponding operating software of the electronic device. It can be a computer or other intelligent terminals such as a smart phone. The embodiment of the present invention does not limit the specific form of the terminal device and can be determined according to actual needs.

[0128] Next, the technical solution of the present invention will be described in detail through specific embodiments. It should be noted that the following specific embodiments can be combined with each other, and the same or similar concepts or processes may not be described again in some embodiments.

[0129] Figure 2a is a schematic flowchart of the first embodiment of the cloth inspection processing method provided by the present invention. As Figure 2a shown, the cloth inspection processing method specifically includes the following steps:

[0130] S201: In response to the user's operation on the start cloth inspection button on the graphical user interface, send a start cloth inspection instruction to the server.

[0131] In this step, when the user needs to perform cloth inspection, the user needs to click the start cloth inspection button on the user graphical interface of the terminal device, and the terminal device will send a start cloth inspection instruction to the server. The start cloth inspection instruction is used to control the cloth inspection machine to start the rotating shaft and start cloth inspection.

[0132] S202: Send a start cloth inspection instruction to the cloth inspection machine.

[0133] In this step, after receiving the start fabric inspection instruction sent by the terminal device, the server sends the start fabric inspection instruction to the fabric inspection machine. After receiving the start fabric inspection instruction, the fabric inspection machine starts the rotating shaft to start fabric inspection.

[0134] After receiving the start fabric inspection instruction sent by the server, the fabric inspection machine controls the rotating shaft of the fabric inspection machine to start rotating through the controller, thereby driving the fabric to be processed to move. At this time, the industrial camera on the fabric inspection machine starts to work, takes pictures of the fabric to be processed, and sends the taken image of the fabric to be processed to the server. At the same time, the encoder on the fabric inspection machine starts to work, collects the cloth length information, and sends a cloth length signal to the server. The cloth length signal contains the cloth length information, and the cloth length information is used by the server to calculate the coordinate information of the defects. Exemplarily, Figure 2b This is a physical diagram of the encoder provided by the present invention, as Figure 2b shown, the encoder can obtain the cloth length information by the rotation of the roller on it. The specific structure and style of the encoder are not limited in the embodiments of the present invention and can be set according to the actual situation.

[0135] It should be noted that the industrial camera taking pictures of the fabric to be processed can be taking videos or taking continuous images. The specific way of taking pictures by the industrial camera is not limited in the embodiments of the present invention and can be set according to the actual situation.

[0136] It should be noted that the image of the fabric to be processed taken by the industrial camera can be each frame image in the video taken by the industrial camera for the fabric to be processed, or can be continuous images taken by the industrial camera, or can be images taken by the industrial camera at a certain time interval. The images are connected to cover the entire fabric. The specific way of obtaining the image of the fabric to be processed from the shooting content of the industrial camera is not specifically limited in the embodiments of the present invention and can be set according to the actual situation.

[0137] S203: Receive the image of the fabric to be processed sent by the fabric inspection machine.

[0138] S204: Input the fabric image into a pre-obtained fabric detection model to determine the defect positions and defect types in the fabric image.

[0139] In the above steps, after receiving the image of the fabric to be processed and the cloth length signal sent by the fabric inspection machine, the server inputs the image of the fabric to be processed into a pre-obtained fabric detection model, and then obtains the defect positions and defect types in the fabric image. After determining the defect positions and defect types, they can be marked.

[0140] It should be noted that the fabric detection model is an image recognition model obtained by training a model using a set of fabric images with calibrated defects and a convolutional network before the fabric inspection operation for identifying the positions and types of fabric defects in the fabric image.

[0141] Exemplarily, Figure 2c The defect type diagram provided by the present invention is shown from left to right as snag, hole, broken warp, and different color. The embodiments of the present invention do not limit the specific types of defects.

[0142] S205: Mark in the cloth image according to the defect position and defect type in the cloth image to obtain the marked cloth image.

[0143] S206: Send the marked cloth image to the user's terminal device.

[0144] In the above steps, after the server determines the defect position and defect type in the cloth image, it can mark the defects so that after sending them to the terminal device, the user can inspect the defects. After the marking is completed, the cloth image and the cloth length information can be sent to the terminal device. The terminal can display the received cloth length information on the user graphical interface to remind the user.

[0145] It should be noted that when there are no defects in the cloth image, there is no need to mark the image, and the image at this time is also called the marked image.

[0146] Exemplarily, Figure 2d The marked cloth image provided by the present invention is shown as Figure 2d shown. In the figure, 5 defects of the roving type and one defect of the stain type and their corresponding positions are marked. It should be noted that Figure 2d only an example of a marked cloth image is given, and the embodiments of the present invention do not specifically limit the defect marking method and defect type, which can be selected according to the actual situation.

[0147] S207: In response to the user's operation on the stop cloth inspection button on the graphical user interface, send a stop cloth inspection instruction to the server.

[0148] In this step, after the terminal device receives the marked cloth image sent by the server, it displays the marked cloth image to the user through the user graphical interface. The user can judge whether the type or position of the marked defect is correct and whether there are any missed defects. If there are no marked defects and no missed defects in the marked cloth image, there is no need to stop cloth inspection. If there are marked defects or missed defects in the marked cloth image, the user needs to click the stop cloth inspection button on the user graphical interface, and the terminal device will send a stop cloth inspection instruction to the server, so that the server can send a stop cloth inspection instruction to the cloth inspection machine to stop the rotating shaft of the cloth inspection machine.

[0149] S208: Send a stop cloth inspection instruction to the cloth inspection machine.

[0150] After the server receives the stop cloth inspection instruction sent by the terminal device, it will send a stop cloth inspection instruction to the cloth inspection machine. After receiving the stop cloth inspection instruction, the cloth inspection machine will stop the rotating shaft.

[0151] After the cloth inspection machine receives the stop cloth inspection instruction sent by the server, it controls the rotating shaft of the cloth inspection machine to stop rotating through the controller, thereby stopping the movement of the cloth.

[0152] The cloth inspection processing method provided by the embodiment of the present invention remotely controls the start of the cloth inspection machine manually. The cloth inspection machine sends the image of the cloth to be processed taken to the server. The server inputs the image into the cloth detection model to obtain the defect position and type and marks them. Then, the marked image is sent to the terminal device. The user can check whether the marks are correct and whether there are any omissions. Then, the user sends a stop cloth inspection instruction to the server through the terminal, and the server then sends the stop cloth inspection instruction to the cloth inspection machine to remotely control the cloth inspection machine to stop cloth inspection. The present invention uses both manual and cloth detection models for cloth inspection, and manual cloth inspection can be performed remotely, effectively reducing the cloth inspection cost and improving the cloth inspection efficiency.

[0153] Figure 3a It is a schematic flowchart of the second embodiment of the cloth inspection processing method provided by the present invention. As Figure 3a shown, on the basis of the above embodiment, after the server sends a stop cloth inspection instruction to the cloth inspection machine, the cloth inspection processing method further includes the following steps:

[0154] S301: In response to the position and type of at least one defect manually marked by the user on the graphical user interface, send defect marking information to the server.

[0155] In this step, after the cloth inspection machine stops the rotating shaft, the user can perform further marking operations on the marked cloth image through the user graphical interface, mark the position and type of the defect, and the terminal sends the defect information to the server so that the server calculates the defect coordinate information according to the defect marking information.

[0156] It should be noted that in the following situations, the user needs to perform further marking operations on the marked cloth image.

[0157] Situation 1: The marked defect position and type in the marked cloth image are both correct and there are no missing defects. At this time, the user's further marking operation is a confirmation marking operation, and the terminal device can send defect marking information to the server according to the confirmation marking operation.

[0158] Situation 2: The marked defect position or type in the marked cloth image is incorrect and there are no missing defects. At this time, the user's further marking operation is to manually mark the correct defect position and type and perform a confirmation marking operation.

[0159] Case 3: The marked defect positions and types in the marked cloth image are correct, but there are missing defects. In this case, the user's further marking operation is to manually mark the positions and types of the missing defects and perform a confirmation marking operation.

[0160] Case 4: The marked defect positions or types in the marked cloth image are incorrect and there are missing defects. In this case, the user's further marking operation is to manually mark the correct defect positions and types and the positions and types of the missing defects and perform a confirmation marking operation.

[0161] Case 5: There are no marked defects in the marked cloth image, but there are missing defects. In this case, the user's further marking operation is to manually mark the positions and types of the missing defects and perform a confirmation marking operation.

[0162] Exemplarily, Figure 3b is a schematic diagram of the user marking defects provided by the present invention. As Figure 3b shown, the user marks a rectangular box at the defect position. Figure 3b This is only an example of a marked cloth image. The present invention embodiments do not specifically limit the defect marking method and defect types, which can be selected according to the actual situation.

[0163] S302: Calculate the coordinate information of at least one defect according to the defect marking message.

[0164] In this step, after the server receives the defect marking information from the terminal device, it can calculate the coordinate information of the defect according to the defect marking information and the cloth length signal sent by the encoder on the cloth inspection machine. Then, the cloth inspection machine can mark the defect according to the coordinate information of the defect.

[0165] The following gives an example of the calculation of the coordinate information of the defect:

[0166] Before the cloth inspection starts, it is necessary to first obtain the scale factor of the industrial camera. According to the pixels of the defect in the image and the scale factor, the abscissa of the defect on the cloth can be calculated. Exemplarily, Figure 3c is a schematic diagram of calculating the scale factor provided by the present invention. As Figure 3c shown, an A4 paper can be pasted on the cloth and photographed by the industrial camera to obtain an image. From the parameters of this image, we can obtain the width of the pixels occupied by the A4 paper, which can be set as w. At the same time, it is known that the actual physical size of the A4 paper is 297 millimeters. Therefore, the scale factor of the camera acquisition system is S = 297 / w, where S represents the scale factor.

[0167] Exemplarily, the position of the defect in the defect marking information is determined in the form of a rectangular frame. Therefore, the central pixel k of this rectangular frame can be calculated, and then the horizontal coordinate X of the defect on the cloth can be calculated as X = k × S. At the same time, by querying the cloth length signal, the number of meters of the cloth length where the defect is located, that is, the vertical coordinate value Y, can be obtained. After obtaining the position coordinates of the defect, the position coordinates XY of the defect are stored in the server.

[0168] It should be noted that the above examples only give an example of the method for obtaining the scale factor, the method for determining the position of the defect, and the method for calculating the coordinate information of the defect. The present invention does not specifically limit the method for obtaining the scale factor, the method for determining the position of the defect, and the method for calculating the coordinate information of the defect, and can be selected according to the actual situation.

[0169] S303: According to the coordinate information of at least one defect, send a marking instruction to the fabric inspection machine.

[0170] In this step, after the server calculates the coordinate information of the defect, a marking instruction can be sent to the fabric inspection machine. The marking instruction is used to control the fabric inspection machine to mark the defect on the cloth, and the marking instruction contains the coordinates and type of the defect.

[0171] After the fabric inspection machine receives the marking instruction, the coding machine on the fabric inspection machine finds the position of the defect according to the horizontal and vertical coordinates of the defect and marks at the position of the defect.

[0172] The fabric inspection processing method provided by the embodiment of the present invention further confirms the defects determined by the fabric detection model, modifies the errors in the defects determined by the fabric detection model, marks the omitted defects, and then sends the defect marking information to the server. Furthermore, the server can calculate the coordinate information of the defect, and the server then sends the coordinate information to the fabric inspection machine, and the fabric inspection machine can mark the cloth according to the coordinate information.

[0173] Figure 3d It is a schematic diagram of the relationship between two fabric inspection methods provided by the present invention. As Figure 3d shown, the present invention adopts a method of combining manual fabric inspection and fabric inspection by a fabric detection model for fabric inspection, with manual fabric inspection as the main and the fabric detection model as the auxiliary, which increases the accuracy of fabric inspection, reduces the cost of fabric inspection, and improves the efficiency of fabric inspection.

[0174] Figure 4a It is a schematic flowchart of the third embodiment of the fabric inspection processing method provided by the present invention. As Figure 4a shown, on the basis of the above embodiment, before sending a start fabric inspection instruction to the server in response to the operation of the start fabric inspection button by the user on the graphical user interface, the fabric inspection processing method further includes the following steps:

[0175] S401: Obtain a pre-calibrated set of cloth images, where the set of cloth images includes multiple calibrated cloth images with the defect positions and defect types of the cloth marked.

[0176] During the cloth inspection process, a pre-obtained cloth detection model is used for cloth inspection. Therefore, before cloth inspection, the cloth detection model needs to be obtained. The cloth detection model is trained based on the pre-calibrated set of cloth images and the convolutional network model.

[0177] In this step, before model training, it is necessary to obtain a pre-calibrated set of cloth images. The set of cloth images includes multiple calibrated cloth images with the defect positions and defect types of the cloth marked. The convolutional neural network can then learn and train based on the defect positions and defect types in the calibrated cloth images.

[0178] It should be noted that the pre-calibrated cloth images can be the cloth images with the defect positions and types marked in the above embodiments, or can be obtained from the network. The embodiments of the present invention do not limit the acquisition method of the pre-calibrated cloth images, and can be selected according to the actual situation.

[0179] Exemplarily, Figure 4b is a schematic diagram of the pre-calibrated cloth image provided by the present invention. As Figure 4b shown, the defect types of holes and stains and the defect positions are marked in this image. It should be noted that Figure 4b is only an example of the pre-calibrated cloth image. The embodiments of the present invention do not specifically limit the calibrated cloth images, and can be selected according to the actual situation.

[0180] S402: According to the set of cloth images, use the convolutional network model for model training to obtain the cloth detection model.

[0181] In this step, after obtaining the set of cloth images, the cloth detection model can be obtained by using the convolutional network model for training. By inputting the cloth image into the cloth detection model, the positions and types of cloth defects can be identified and determined.

[0182] The following is an example of the training process of the cloth detection model:

[0183] The set of cloth images contains at least 100 labeled cloth images for each type of defect. Before model training, it is necessary to convert the labeled cloth images into labeled files.

[0184] Exemplarily, the following is an example of the labeled file:

[0185]

[0186]

[0187] The above code is an example of an annotation file. In this annotation file, the width of the cloth image is marked as 800 pixels, the height is 600 pixels, there are two defects, the type of the defect is a hole. The minimum abscissa of the rectangular box where one defect is located is 353 pixels, the maximum ordinate is 439 pixels, and the minimum abscissa of the rectangular box where the other defect is located is 383 pixels, and the maximum ordinate is 599 pixels.

[0188] It should be noted that the above example is only an example of the annotation file, and the embodiments of the present invention do not limit the annotation file, which can be set according to the actual situation.

[0189] Exemplarily, the convolutional network model is selected as the YOLO network, the annotation file is imported into the YOLO network for iterative training, and the iteration is at least 2000 times until the convergence descent rate is lower than 0.001, then the training can be stopped to obtain a trained cloth detection model. Input the cloth image into the cloth detection model, and the position and type of the defect can be obtained. It should be noted that the embodiments of the present invention do not specifically limit the convolutional network model and the training process, which can be selected according to the actual situation.

[0190] It should be noted that the above training process is only an example of the training process of the cloth detection model, and the embodiments of the present invention do not specifically limit the training process of the cloth detection model, which can be set according to the actual situation.

[0191] The cloth inspection processing method provided by the embodiments of the present invention trains a cloth detection model through a pre-calibrated cloth image and a convolutional network model, and then performs cloth inspection through the cloth detection model, which can reduce the cloth inspection cost and improve the cloth inspection efficiency.

[0192] The following is an embodiment of the device of the present invention, which can be used to execute the method embodiment of the present invention. For the details not disclosed in the embodiment of the device of the present invention, please refer to the method embodiment of the present invention.

[0193] Figure 5 It is a schematic structural diagram of the first embodiment of the cloth inspection processing device provided by the present invention; as Figure 5 shown, the cloth inspection processing device 50 includes:

[0194] A receiving module 51, configured to receive a cloth image to be processed sent by a cloth inspection machine during the cloth inspection process;

[0195] A detection module 52, configured to input the cloth image into a pre-acquired cloth detection model to determine the defect position and defect type in the cloth image, and the cloth detection model is used to identify the position and type of the cloth defect in the cloth image;

[0196] A processing module 53, configured to mark in the cloth image according to the defect positions and defect types in the cloth image, so as to obtain a marked cloth image;

[0197] A sending module 54, configured to send the marked cloth image to the user's terminal device;

[0198] The sending module 54 is further configured to send the stop cloth inspection instruction to the cloth inspection machine after receiving the stop cloth inspection instruction sent by the terminal device, where the stop cloth inspection instruction is an instruction generated by the terminal device in response to the user's operation and used to control the cloth inspection machine to stop the rotating shaft.

[0199] Further, the receiving module 51 is further configured to receive the start cloth inspection instruction sent by the terminal device.

[0200] Further, the sending module 54 is further configured to send the start cloth inspection instruction to the cloth inspection machine.

[0201] Further, the receiving module 51 is further configured to obtain a pre-calibrated cloth image set, where the cloth image set includes a plurality of calibrated cloth images with marked defect positions and defect types of the cloth.

[0202] Further, the processing module 53 is further configured to perform model training using a convolutional network model according to the cloth image set to obtain the cloth detection model.

[0203] Further, the receiving module 51 is further configured to receive a defect marking message sent by the terminal device, where the defect marking message includes the positions and types of at least one defect manually marked by the user.

[0204] Further, the processing module 53 is further configured to calculate the coordinate information of the at least one defect according to the defect marking message.

[0205] Further, the sending module 54 is further configured to send a marking instruction to the cloth inspection machine according to the coordinate information of the at least one defect, where the marking instruction includes the coordinate information of the at least one defect, and the marking instruction is used to control the cloth inspection machine to mark the at least one defect on the cloth.

[0206] Further, the receiving module 51 is further configured to receive cloth inspection requirement information sent by the factory cloth inspection platform, where the cloth inspection requirement information includes the rating of the required cloth inspection users and the number of cloth inspection users.

[0207] Further, the processing module 53 is further configured to obtain at least one target cloth inspection user that meets the cloth inspection requirement information from the currently selectable cloth inspection users according to the pre-obtained rating information of each user and the cloth inspection requirement information.

[0208] Further, the processing module 53 is further configured to send a cloth inspection task to the terminal device corresponding to at least one target cloth inspection user.

[0209] In a possible design of the embodiment of the present invention, the rating information of each cloth inspection user includes the rating of the cloth inspection user, and the video playback speed, false negative rate, and false positive rate corresponding to the rating; the rating information is obtained based on the cloth inspection data of the cloth inspection user during the cloth inspection training process.

[0210] The cloth inspection processing device provided in this embodiment is used to execute the technical solution of the server in any of the foregoing method embodiments. The implementation principle and technical effects are similar. By connecting the server with the user's terminal device and the cloth inspection machine, manual remote cloth inspection can be realized, and the cloth inspection model is used in the server for cloth inspection, effectively reducing the cloth inspection cost and improving the cloth inspection efficiency.

[0211] Figure 6a It is a schematic structural diagram of the second embodiment of the cloth inspection processing device provided by the present invention; as Figure 6a shown, the cloth inspection processing device 60 includes:

[0212] A receiving module 61, configured to receive the marked cloth image sent by the server during the cloth inspection process, where the marked cloth image includes the defect positions and defect types automatically detected and marked by the server;

[0213] A display module 62, configured to display the marked cloth image on the graphical user interface;

[0214] A sending module 63, configured to send a stop cloth inspection instruction to the server in response to the user's operation on the stop cloth inspection button on the graphical user interface, where the stop cloth inspection instruction is used to control the cloth inspection machine to stop the rotating shaft.

[0215] Further, the sending module 63 is further configured to:

[0216] In response to the user's operation on the start cloth inspection button on the graphical user interface, send a start cloth inspection instruction to the server, where the start cloth inspection instruction is used to control the cloth inspection machine to start the rotating shaft and start cloth inspection.

[0217] Further, the sending module 63 is further configured to:

[0218] In response to the position and type of at least one defect manually marked by the user on the graphical user interface, send defect marking information to the server, where the defect marking information includes the position and type of the at least one defect.

[0219] Further, the receiving module 61 is further configured to receive the cloth inspection task sent by the server and display the cloth inspection task on the graphical user interface.

[0220] Figure 6b Schematic structural diagram of the third embodiment of the cloth inspection processing device provided by the present invention; as Figure 6b shown, the cloth inspection processing device 60 further includes:

[0221] A processing module 64, configured to obtain cloth inspection data in response to an operation of the user on the graphical user interface during the cloth inspection training process.

[0222] Further, the sending module 63 is further configured to send the cloth inspection data to the server, and the cloth inspection data is used to obtain the rating information of the user. The rating information includes the rating of the user, and the video playback speed, false negative rate, and false positive rate corresponding to the rating.

[0223] The cloth inspection processing device provided in this embodiment is used to execute the technical solutions of the user's terminal device in any of the foregoing method embodiments. The implementation principles and technical effects are similar. The user can perform cloth inspection through the terminal device, and can also check the defect positions and defect types determined by the cloth detection model, effectively reducing the cloth inspection cost and improving the cloth inspection efficiency.

[0224] Figure 7 Schematic structural diagram of a server provided by the present invention. As Figure 7 shown, the server 70 includes:

[0225] A processor 71, a memory 72, and a communication interface 73;

[0226] The memory 72 is used to store executable instructions of the processor 71;

[0227] Wherein, the processor 71 is configured to execute the technical solutions of the server in any of the foregoing method embodiments by executing the executable instructions.

[0228] Optionally, the memory 72 can be either independent or integrated with the processor 71.

[0229] Optionally, when the memory 72 is a device independent of the processor 71, the server 70 may further include:

[0230] A bus for connecting the above-mentioned devices.

[0231] This server is used to execute the technical solutions on the server side in any of the foregoing method embodiments. The implementation principles and technical effects are similar, and will not be elaborated here.

[0232] Figure 8 A structural schematic diagram of a terminal device provided by the present invention. As Figure 8 shown, the service terminal device 80 includes:

[0233] a processor 81, a memory 82, a display 83, and a communication interface 84;

[0234] The memory 82 is used to store executable instructions of the processor;

[0235] Wherein, the processor 81 is configured to execute the technical solutions on the terminal device side in any of the foregoing embodiments by executing the executable instructions.

[0236] Optionally, the memory 82 can be either independent or integrated with the processor 81.

[0237] Optionally, when the memory 82 is a device independent of the processor 81, the terminal device 80 may further include:

[0238] a bus for connecting the above-mentioned devices.

[0239] This terminal device is used to execute the technical solutions of the terminal device in any of the foregoing method embodiments, and its implementation principle and technical effects are similar, which will not be elaborated here.

[0240] The embodiment of the present invention also provides a readable storage medium, on which a computer program is stored. When the computer program is executed by a processor, it implements the technical solutions on the server or terminal device side provided in any of the foregoing embodiments.

[0241] The embodiment of the present invention also provides a computer program product, including a computer program, which is used to implement the technical solutions on the server or terminal device side provided in any of the foregoing method embodiments when executed by a processor.

[0242] Those of ordinary skill in the art can understand that all or part of the steps of implementing the foregoing method embodiments can be completed by hardware related to program instructions. The foregoing program can be stored in a computer-readable storage medium. When the program is executed, it executes the steps including the foregoing method embodiments; and the foregoing storage medium includes: various media such as ROM, RAM, magnetic disk, or optical disc that can store program codes.

[0243] Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention, rather than to limit it; although the present invention has been described in detail with reference to the foregoing embodiments, those of ordinary skill in the art should understand that they can still modify the technical solutions described in the foregoing embodiments, or perform equivalent replacements on some or all of the technical features; and these modifications or replacements do not cause the essence of the corresponding technical solutions to deviate from the scope of the technical solutions of the embodiments of the present invention.

Claims

1. A cloth inspection processing method, characterized in that, applied to a server, the method includes: Receiving cloth inspection requirement information sent by a factory cloth inspection platform, where the cloth inspection requirement information includes the ratings of cloth inspection users required and the number of cloth inspection users; According to the pre-acquired rating information of each user and the cloth inspection requirement information, obtaining at least one target cloth inspection user that meets the cloth inspection requirement information from the currently selectable cloth inspection users; Sending a cloth inspection task to the terminal devices corresponding to the at least one target cloth inspection user; Among them, the rating information of each cloth inspection user includes the rating of the cloth inspection user, and the video playback speed, false negative rate and false positive rate corresponding to the rating; the rating information is obtained based on the cloth inspection data of the cloth inspection user during the cloth inspection training process; During the cloth inspection process, receiving the cloth image to be processed sent by the cloth inspection machine; Inputting the cloth image into a pre-acquired cloth detection model to determine the defect position and defect type in the cloth image, where the cloth detection model is used to identify the position and type of cloth defects in the cloth image; According to the defect position and defect type in the cloth image, marking in the cloth image to obtain a marked cloth image, and sending the marked cloth image to the terminal devices of the at least one target cloth inspection user; After receiving the stop cloth inspection instruction sent by the terminal device, sending the stop cloth inspection instruction to the cloth inspection machine, where the stop cloth inspection instruction is an instruction generated by the terminal device in response to the operation of the at least one target cloth inspection user for controlling the cloth inspection machine to stop rotating the shaft, and the stop cloth inspection instruction is used to indicate that there are marked defects or missed defects in the marked cloth image; Receiving the defect marking message sent by the terminal device, where the defect marking message includes the positions and types of at least one defect manually marked by the at least one target cloth inspection user, and the defect marking information is manually marked by the at least one target cloth inspection user on the graphical interface; Calculating the coordinate information of the at least one defect according to the defect marking message and the cloth length signal sent by the encoder on the cloth inspection machine; According to the coordinate information of the at least one defect, sending a marking instruction to the cloth inspection machine, where the marking instruction includes the coordinate information of the at least one defect, and the marking instruction is used to control the cloth inspection machine to mark the at least one defect on the cloth.

2. The method according to claim 1, characterized in that, Before receiving the cloth image to be processed sent by the cloth inspection machine during the cloth inspection process, the method further includes: Receiving the start cloth inspection instruction sent by the terminal device and sending the start cloth inspection instruction to the cloth inspection machine, where the start cloth inspection instruction is used to control the cloth inspection machine to start rotating the shaft and start cloth inspection.

3. The method according to claim 1 or 2, characterized in that, The method further includes: Obtaining a pre-calibrated set of cloth images, where the set of cloth images includes a plurality of calibrated cloth images with the defect positions and defect types of the marked cloth. Based on the set of cloth images, a convolution network model is used for model training to obtain the cloth detection model.

4. A cloth inspection processing method characterized in that it is applied to a terminal device, and the method includes: Receiving a cloth inspection task sent by a server and displaying the cloth inspection task on a graphical user interface; During the cloth inspection process, receiving the marked cloth images sent by the server, where the marked cloth images include the defect positions and defect types automatically detected and marked by the server; Displaying the marked cloth images on the graphical user interface; In response to the operation of at least one target cloth inspection user on the stop cloth inspection button on the graphical user interface, sending a stop cloth inspection instruction to the server, where the stop cloth inspection instruction is an instruction for controlling the cloth inspection machine to stop the rotating shaft, and the stop cloth inspection instruction is used to indicate that there are marked defects or missed defects in the marked cloth images. The at least one target cloth inspection user is determined based on cloth inspection requirement information and the rating information of each user. The cloth inspection requirement information includes the rating of the required cloth inspection users and the number of cloth inspection users. The at least one target cloth inspection user meets the cloth inspection requirement information. The rating information of each cloth inspection user includes the rating of the cloth inspection user and the video playback speed, false negative rate, and false positive rate corresponding to the rating. The rating information is obtained based on the cloth inspection data of the cloth inspection user during the cloth inspection training process; Sending a defect marking message to the server, where the defect marking message includes the positions and types of at least one defect manually marked by the at least one target cloth inspection user, and the defect marking information is manually marked by the at least one target cloth inspection user on the graphical interface; wherein the defect marking information is used by the server to calculate the coordinate information of the at least one defect based on the defect marking message and the cloth length signal sent by the encoder on the cloth inspection machine, and based on the coordinate information of the at least one defect, sending a marking instruction to the cloth inspection machine, where the marking instruction includes the coordinate information of the at least one defect, and the marking instruction is used to control the cloth inspection machine to mark the at least one defect on the cloth.

5. The method according to claim 4 characterized in that before receiving the marked cloth images sent by the server during the cloth inspection process, the method further includes: In response to the operation of at least one target cloth inspection user on the start cloth inspection button on the graphical user interface, sending a start cloth inspection instruction to the server, where the start cloth inspection instruction is used to control the cloth inspection machine to start the rotating shaft and start cloth inspection.

6. The method according to claim 4 or 5 characterized in that the method further includes: In response to the positions and types of at least one defect manually marked by the at least one target cloth inspection user on the graphical user interface, sending defect marking information to the server, where the defect marking information includes the positions and types of the at least one defect.

7. The method according to claim 4 or 5 characterized in that the method further includes: Receive the cloth inspection task sent by the server and display the cloth inspection task on the graphical user interface.

8. The method according to claim 7, wherein, the method further includes: During the cloth inspection training, in response to the operation of the at least one target cloth inspection user on the graphical user interface, obtain cloth inspection data; Send the cloth inspection data to the server, where the cloth inspection data is used to obtain the rating information of the at least one target cloth inspection user, and the rating information includes the rating of the user, and the video playback speed, false negative rate, and false positive rate corresponding to the rating.

9. A cloth inspection processing device, wherein, comprising: A receiving module, configured to receive cloth inspection requirement information sent by a factory cloth inspection platform, where the cloth inspection requirement information includes the rating of the required cloth inspection user and the number of cloth inspection users; According to the pre-obtained rating information of each user and the cloth inspection requirement information, obtain at least one target cloth inspection user that meets the cloth inspection requirement information from the currently selectable cloth inspection users; Send a cloth inspection task to the terminal device corresponding to the at least one target cloth inspection user; Wherein, the rating information of each cloth inspection user includes the rating of the cloth inspection user, and the video playback speed, false negative rate, and false positive rate corresponding to the rating; the rating information is obtained according to the cloth inspection data of the cloth inspection user during the cloth inspection training process; During the cloth inspection process, receive the cloth image to be processed sent by the cloth inspection machine; A detection module, configured to input the cloth image into a pre-obtained cloth detection model to determine the defect position and defect type in the cloth image, where the cloth detection model is used to identify the position and type of cloth defects in the cloth image; A processing module, configured to mark the defect position and defect type in the cloth image in the cloth image to obtain a marked cloth image; A sending module, configured to send the marked cloth image to the terminal device of at least one target cloth inspection user; The sending module is further configured to, after receiving the stop cloth inspection instruction sent by the terminal device, send the stop cloth inspection instruction to the cloth inspection machine, where the stop cloth inspection instruction is an instruction generated by the terminal device in response to the operation of the at least one target cloth inspection user for controlling the cloth inspection machine to stop rotating the shaft, and the stop cloth inspection instruction is used to indicate that there are marked defects or omitted defects in the marked cloth image; Receive the defect marking message sent by the terminal device, where the defect marking message includes the position and type of at least one defect manually marked by the at least one target cloth inspection user, and the defect marking information is manually marked by the at least one target cloth inspection user on the graphical interface; Calculate the coordinate information of the at least one defect according to the defect marking message and the cloth length signal sent by the encoder on the cloth inspection machine; According to the coordinate information of the at least one defect, send a marking instruction to the cloth inspection machine, where the marking instruction includes the coordinate information of the at least one defect, and the marking instruction is used to control the cloth inspection machine to mark the at least one defect on the cloth.

10. A cloth inspection processing device, characterized in that, it includes: a receiving module, configured to receive the cloth inspection task sent by the server and display the cloth inspection task on the graphical user interface; During the cloth inspection process, receive the marked cloth image sent by the server, and the marked cloth image includes the defect positions and defect types automatically detected and marked by the server; a display module, configured to display the marked cloth image on the graphical user interface; a sending module, configured to send a stop cloth inspection instruction to the server in response to the operation of at least one target cloth inspection user on the stop cloth inspection button on the graphical user interface, where the stop cloth inspection instruction is used to control the cloth inspection machine to stop the rotation shaft, and the stop cloth inspection instruction is used to indicate that there are marked defects or missing defects in the marked cloth image. The at least one target cloth inspection user is determined based on the cloth inspection requirement information and the rating information of each user. The cloth inspection requirement information includes the rating of the required cloth inspection users and the number of cloth inspection users. The at least one target cloth inspection user meets the cloth inspection requirement information. The rating information of each cloth inspection user includes the rating of the cloth inspection user and the video playback speed, false negative rate, and false positive rate corresponding to the rating; The rating information is obtained based on the cloth inspection data of the cloth inspection user during the cloth inspection training process; send a defect marking message to the server, where the defect marking message includes the positions and types of at least one defect manually marked by the at least one target cloth inspection user, and the defect marking information is manually marked by the at least one target cloth inspection user on the graphical interface; wherein, the defect marking information is used for the server to calculate the coordinate information of the at least one defect according to the defect marking message and the cloth length signal sent by the encoder on the cloth inspection machine, and send a marking instruction to the cloth inspection machine according to the coordinate information of the at least one defect. The marking instruction includes the coordinate information of the at least one defect, and the marking instruction is used to control the cloth inspection machine to mark the at least one defect on the cloth.

11. A server, characterized in that, it includes: a processor, a memory, and a communication interface; The memory is used to store the executable instructions of the processor; wherein, the processor is configured to execute the cloth inspection processing method according to any one of claims 1 to 3 by executing the executable instructions.

12. A terminal device, characterized in that, it includes: a processor, a memory, a display, and a communication interface; The memory is used to store the executable instructions of the processor; wherein, the processor is configured to execute the cloth inspection processing method according to any one of claims 4 to 8 by executing the executable instructions.

13. A cloth inspection processing system, characterized in that, it includes: a server, a terminal device, a switch, and a cloth inspection machine; The server is used to execute the cloth inspection processing method according to any one of claims 1 to 3; The terminal device is used to execute the cloth inspection processing method according to any one of claims 4 to 8; The switch is used to transmit network communication signals; The fabric inspection machine is used to receive the start fabric inspection instruction, the stop fabric inspection instruction, and the marking instruction sent by the server, control the rotation shaft on the fabric inspection machine to start rotating according to the start fabric inspection instruction, control the rotation shaft on the fabric inspection machine to stop rotating according to the stop fabric inspection instruction, and control the coder on the fabric inspection machine to mark the defects on the fabric according to the marking instruction; The fabric inspection machine is further used to send the image of the fabric to be processed to the server.

14. A readable storage medium, on which a computer program is stored, characterized in that, when the computer program is executed by a processor, it implements the fabric inspection processing method according to any one of claims 1 to 8.

Images