Cellulose diacetate tow appearance quality detection method and system

Abstract

The invention discloses a diacetate fiber tow appearance quality detection method and system. The method comprises the following steps: S1, under an illumination condition covering a detection range, performing image acquisition on a moving fiber tow, performing conversion processing, and uploading the image to a computer; s2, measuring the broadening of the fiber tow, and further detecting the broadband fluctuation condition; and S3, detecting the defects including slub yarns and jump rolls by using a target detection algorithm model so as to identify and classify various defects. The invention provides an efficient, reliable and intelligent fiber tow quality detection method, various limitations in a traditional detection method are overcome by integrating the AI technology, and the overall quality of the diacetate fiber tow is remarkably improved.

Description

Technical Field

[0001] The invention relates to the field of image detection and analysis, and in particular to a method and system for detecting the appearance quality of a diacetate fiber tow. Background Art

[0002] The appearance defects of cellulose acetate tow for cigarettes mainly include skipped curls, bamboo knots, splitting, large differences in curl ratios, and large standard deviations (Std) of ribbon width. These defects not only affect the appearance of the tow, but also its quality. In addition, since the ribbon becomes an elastic body after curling, the appearance of the tow may vary significantly under different pre-tension conditions.

[0003] Patent No. 201210049619.2 discloses a tow appearance quality inspection system, which can detect the quality problems of the tow in real time under the condition of high-speed operation of the tow. The system includes a dark box, a white pad, an ARM controller, a camera, a flicker-free LED light source, a power controller, a video conversion module, a PC and an alarm. The camera collects the video signal of the tow in real time, the ARM controller determines whether the tow has quality problems through an integrated algorithm, the PC records and displays the video signal, and sounds an alarm through the alarm when a defect is detected. However, in the actual tow production process, the running speed of the tow is as high as 10m / s, and there are jitter and jump problems in the running trajectory, which makes it difficult to capture high-definition images. This not only brings huge challenges to the subsequent algorithm writing and stable operation of the equipment, but also seriously affects the detection rate of surface defects of the tow. Summary of the invention

[0004] In order to solve the existing problems, the present invention significantly improves the recognition rate of defects such as slubby yarn and skipped coils by optimizing the YOLO V8 algorithm and applying data enhancement technology. It is particularly suitable for the appearance quality inspection of diacetate fiber tow under complex motion conditions. The specific scheme is as follows:

[0005] A method for detecting the appearance quality of a diacetate fiber tow comprises the following steps:

[0006] S1, under the lighting conditions covering the detection range, the image of the moving fiber bundle is collected and converted and then uploaded to the computer;

[0007] S2, measuring the fiber bundle width and further detecting the broadband fluctuation;

[0008] S3, uses the target detection algorithm model to detect defects including bamboo knots and skipped coils, thereby identifying and classifying various defects.

[0009] Preferably, step S2 specifically includes the following steps:

[0010] S21, coarse positioning of the fiber bundle;

[0011] S22, uses sub-pixel edge detection operators for precise positioning, extracts fiber width edges, and improves measurement accuracy. By analyzing the changes in pixel intensity near the edge and using interpolation or fitting methods, the exact position of the edge can be more accurately estimated, thereby capturing the subtle boundaries of the fiber material, ensuring that the measurement results are more reliable and accurate, and thus optimizing the detection effect of the fiber width;

[0012] S23, converting pixels to actual physical sizes through pixel equivalent calibration, calculating width data, and further detecting width fluctuations.

[0013] Preferably, step S21 of roughly positioning the fiber bundle specifically includes the following method steps: performing rough positioning in sequence through distortion correction, image filtering, image enhancement, traditional edge detection operators and morphological processing.

[0014] Preferably, the target detection algorithm used in step S3 is YOLO V8.

[0015] Preferably, the step of performing detection using the target detection algorithm model in step S3 specifically includes:

[0016] S31, data augmentation is performed to expand the data set, including applying a series of transformations such as rotation, flipping, scaling, cropping, color change, etc. to the original data, thereby effectively expanding the scale of the training data;

[0017] S32, dividing the enhanced data into a training set and a test set, and manually annotating them;

[0018] S33, modify the hyperparameters and perform iterative training;

[0019] S34, selecting the optimal training model and performing test set testing;

[0020] S35, the target detection algorithm model uses deep learning algorithms to perform high-definition image acquisition and real-time analysis under high-speed operation and complex motion trajectories of the tow, and automatically identifies and classifies various defects.

[0021] The present invention also discloses a computer-readable storage medium, on which a computer program is stored. After the computer program is run, any of the above methods is executed.

[0022] The present invention also discloses a computer system, including a processor and a storage medium, wherein a computer program is stored on the storage medium, and the processor reads and runs the computer program from the storage medium to execute any of the methods described above.

[0023] Preferably, the system based on any of the above methods comprises an image acquisition module, an illumination module, a ribbon width detection module, and other defect detection modules;

[0024] The image acquisition module is used to acquire high-resolution images of the moving fiber bundles under well-controlled lighting conditions, and convert the acquired image signals into digital signals, which are uploaded to the computer for further analysis by the ribbon width detection module and other defect detection modules;

[0025] The lighting module highlights the features of the object under test through reasonable lighting design, adopts a white strip LED light source that matches the shape of the object under test, ensures uniform lighting and reduces the impact of ambient light on image quality, and ensures sufficient contrast between the image background and the object under test;

[0026] The ribbon width detection module is used to calculate width data and further detect width fluctuations;

[0027] The other defect detection module uses the YOLO V8 target detection algorithm to perform high-definition image acquisition and real-time analysis to automatically identify and classify various defects.

[0028] The beneficial effects of the present invention are:

[0029] The present invention can realize high-definition image acquisition and real-time analysis under operating conditions by using deep learning algorithms, greatly improving detection accuracy. It can process large amounts of data in real time, quickly identify and classify a variety of defects, and significantly improve detection efficiency. Through continuous optimization of machine learning, the detection rate and accuracy are continuously improved, reducing the possibility of missed detection and false detection, and ensuring the stable operation of the system. It can also automatically adapt to different operating conditions, reducing dependence on manual intervention, and improving the automation and intelligence level of the detection process. BRIEF DESCRIPTION OF THE DRAWINGS

[0030] In order to more clearly illustrate the embodiments of the present invention or the technical solutions in the prior art, the following briefly introduces the drawings required for use in the embodiments or the description of the prior art. Obviously, the drawings described below are some embodiments of the present invention. For ordinary technicians in this field, other drawings can be obtained based on these drawings without paying creative work.

[0031] Figure 1 A schematic diagram of the system arrangement of the present invention;

[0032] Figure 2 This is a training flow chart for the present invention when using the YOLO V8 model for detection;

[0033] Figure 3 It is a flow chart of sub-pixel edge detection operator detection;

[0034] Figure 4 Flowchart for pixel equivalent calibration. DETAILED DESCRIPTION

[0035] In order to make the purpose, technical solution and advantages of the embodiments of the present invention clearer, the technical solution in the embodiments of the present invention will be clearly and completely described below in conjunction with the drawings in the embodiments of the present invention. Obviously, the described embodiments are part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of the present invention.

[0036] A method for detecting the appearance quality of a diacetate fiber tow comprises the following steps:

[0037] S1, under well-controlled lighting conditions, captures images of moving fiber bundles, pre-processes the captured images - including denoising, enhancement, etc., and then uploads them to the computer for subsequent analysis.

[0038] S2, measuring the fiber bundle width, that is, measuring the fiber bundle width uniformity, and further detecting the broadband fluctuation. Specifically comprising the following steps:

[0039] S21, coarsely positioning the fiber bundle through distortion correction, image filtering, image enhancement, traditional edge detection operator and morphological processing in sequence.

[0040] S22 uses sub-pixel edge detection operators for precise positioning, extracts fiber width edges, and improves measurement accuracy. By analyzing the changes in pixel intensity near the edge and using interpolation or fitting methods, the exact position of the edge can be more accurately estimated, thereby capturing the subtle boundaries of the fiber material, ensuring that the measurement results are more reliable and accurate, and thus optimizing the detection effect of the fiber width.

[0041] Specifically, Figure 3 ——Sub-pixel edge detection operator detection flow chart, which describes the following steps in detail: First, input the image and perform distortion correction, then remove noise and enhance the image quality through image filtering. Then use the traditional edge detection method to preliminarily detect the edges in the image. On this basis, the edges are refined at the sub-pixel level to improve the accuracy of edge positioning. Subsequently, morphological processing is used to further optimize and improve the edges, and finally output accurate edge detection results.

[0042] S23, converting pixels to actual physical sizes through pixel equivalent calibration, calculating width data, and further detecting width fluctuations.

[0043] Specifically, Figure 4 ——The flowchart of pixel equivalent calibration describes the following steps in detail: First, load the input image and detect the calibration plate in the image. Next, extract the corner point data from the calibration plate. Based on these corner point data, calculate the intrinsic parameters of the camera. Then, use the camera intrinsic parameters and the actual physical size to calculate the actual unit length (pixel equivalent) corresponding to each pixel. Finally, output the calibration result, including detailed data of pixel equivalent and camera intrinsic parameters.

[0044] S3, the defects including bamboo yarn and skipped coils are detected using the target detection algorithm model, so as to identify and classify various defects. Specifically, the target detection algorithm used is YOLO V8, and the model training flow chart is as follows: Figure 2 , describes in detail the steps of data enhancement, training set division, hyperparameter adjustment, model training and testing. The specific steps include:

[0045] S31, data augmentation is performed to expand the data set, including applying a series of transformations such as rotation, flipping, scaling, cropping, color change, etc. to the original data, thereby effectively expanding the scale of the training data.

[0046] S32, dividing the enhanced data into a training set and a test set, and manually annotating them.

[0047] S33, adjust hyperparameters and perform iterative training to optimize model performance.

[0048] S34, select the optimal training model and perform test set testing.

[0049] S35, the target detection algorithm model uses deep learning algorithms to perform high-definition image acquisition and real-time analysis under high-speed operation and complex motion trajectories of the tow, and automatically identifies and classifies various defects.

[0050] The present invention also discloses a computer-readable storage medium, on which a computer program is stored. After the computer program is run, any of the above methods is executed.

[0051] The present invention also discloses a computer system, including a processor and a storage medium, wherein a computer program is stored on the storage medium, and the processor reads and runs the computer program from the storage medium to execute any of the methods described above.

[0052] like Figure 1 , a system based on any of the above methods includes an image acquisition module, a lighting module, and a computer-side AI detection system, wherein the computer-side AI detection system includes a ribbon width detection module and other defect detection modules.

[0053] The image acquisition module is used to acquire images of the moving fiber bundles and realize the conversion of image signals, and upload the images to the computer for use by the ribbon broadband detection module and other defect detection modules.

[0054] The lighting module is used to highlight the detection features of the object to be measured. The reasonable design of the lighting module has an important influence on the image quality and subsequent processing effects. Reasonable lighting of the object to be measured can overcome the influence of ambient light, make the background faded and uniform, highlight the detection features of the object to be measured, ensure that the object and the background have a large contrast, clear boundaries, and maintain stable image quality, reducing the complexity of image processing. When selecting an LED light source, the shape and color of the light source need to be considered. Usually, a light source with the same shape as the object to be measured is selected. The fiber bundle is rectangular, so a white strip LED light source is selected, and the size is determined to be 300mm to cover the detection range.

[0055] The ribbon width detection module is used to calculate width data and further detect width fluctuations.

[0056] The other defect detection module uses the YOLO V8 target detection algorithm to perform high-definition image acquisition and real-time analysis to automatically identify and classify various defects.

[0057] The present invention can realize high-definition image acquisition and real-time analysis under operating conditions by using deep learning algorithms, greatly improving detection accuracy. It can process large amounts of data in real time, quickly identify and classify a variety of defects, and significantly improve detection efficiency. Through continuous optimization of machine learning, the detection rate and accuracy are continuously improved, reducing the possibility of missed detection and false detection, and ensuring the stable operation of the system. It can also automatically adapt to different operating conditions, reducing dependence on manual intervention, and improving the automation and intelligence level of the detection process.

[0058] Those skilled in the art will further appreciate that the various illustrative logic blocks, modules, circuits, and algorithm steps described in conjunction with the embodiments disclosed herein may be implemented as electronic hardware, computer software, or a combination of the two. To clearly illustrate this interchangeability of hardware and software, various illustrative components, blocks, modules, circuits, and steps are generally described above in terms of their functionality. Whether such functionality is implemented as hardware or software depends on the specific application and the design constraints imposed on the overall system. The technician may implement the described functionality in different ways for each specific application, but such implementation decisions should not be interpreted as resulting in a departure from the scope of the present invention.

[0059] The various illustrative logic blocks, modules, and circuits described in conjunction with the embodiments disclosed herein may be implemented or performed with a general purpose processor, a digital signal processor (DSP), an application specific integrated circuit (ASIC), a field programmable gate array (FPGA) or other programmable logic device, discrete gate or transistor logic, discrete hardware components, or any combination thereof designed to perform the functions described herein. A general purpose processor may be a microprocessor, but in the alternative, the processor may be any conventional processor, controller, microcontroller, or state machine. The processor may also be implemented as a combination of computing devices, such as a combination of a DSP and a microprocessor, a plurality of microprocessors, one or more microprocessors in cooperation with a DSP core, or any other such configuration.

[0060] The steps of the method or algorithm described in conjunction with the embodiments disclosed herein may be embodied directly in hardware, in a software module executed by a processor, or in a combination of the two. The software module may reside in a RAM memory, a flash memory, a ROM memory, an EPROM memory, an EEPROM memory, a register, a hard disk, a removable disk, a CD-ROM, or any other form of storage medium known in the art. An exemplary storage medium is coupled to a processor so that the processor can read and write information from / to the storage medium. In an alternative, a storage medium may be integrated into a processor. The processor and the storage medium may reside in an ASIC. The ASIC may reside in a user terminal. In an alternative, the processor and the storage medium may reside in a user terminal as discrete components.

[0061] In one or more exemplary embodiments, the functions described may be implemented in hardware, software, firmware, or any combination thereof. If implemented as a computer program product in software, each function may be stored on or transmitted by a computer-readable medium as one or more instructions or codes. Computer-readable media include both computer storage media and communication media, including any medium that facilitates the transfer of a computer program from one place to another. Storage media may be any available medium that can be accessed by a computer. As an example and not limitation, such a computer-readable medium may include RAM, ROM, EEPROM, CD-ROM or other optical disk storage, disk storage or other magnetic storage device, or any other medium that can be used to carry or store the desired program code in the form of an instruction or data structure and can be accessed by a computer. Any connection is also properly referred to as a computer-readable medium. For example, if the software is transmitted from a website, a server, or other remote source using a coaxial cable, a fiber optic cable, a twisted pair, a digital subscriber line (DSL), or wireless technologies such as infrared, radio, and microwaves, the coaxial cable, fiber optic cable, twisted pair, DSL, or wireless technologies such as infrared, radio, and microwaves are included in the definition of the medium. Disk and disc as used herein include compact disc (CD), laser disc, optical disc, digital versatile disc (DVD), floppy disk and Blu-ray disc, wherein disk often reproduces data magnetically, while disc reproduces data optically with lasers. Combinations of the above should also be included within the scope of computer-readable media.

[0062] The previous description of the disclosure is provided to enable any person skilled in the art to make or use the disclosure. Various modifications to the disclosure will be apparent to those skilled in the art, and the general principles defined herein may be applied to other variations without departing from the spirit or scope of the disclosure. Thus, the disclosure is not intended to be limited to the examples and designs described herein, but should be granted the widest scope consistent with the principles and novel features disclosed herein.

[0063] Although the present invention has been described in detail with reference to the aforementioned embodiments, those skilled in the art should understand that they can still modify the technical solutions described in the aforementioned embodiments, or make equivalent substitutions for some of the technical features therein; and these modifications or substitutions do not make the essence of the corresponding technical solutions deviate from the spirit and scope of the technical solutions of the embodiments of the present invention.

Claims

1. A method for detecting the appearance quality of diacetate fiber tow, characterized in that: The following steps are involved: S1, under the lighting conditions covering the detection range, the image of the moving fiber bundle is collected and converted and then uploaded to the computer; S2, measuring the fiber bundle width and further detecting the broadband fluctuation; S3, uses the target detection algorithm model to detect defects including bamboo knots and skipped coils, thereby identifying and classifying various defects.

2. The method according to claim 1, characterized in that: Step S2 specifically includes the following steps: S21, coarse positioning of the fiber bundle; S22, uses sub-pixel edge detection operators for precise positioning, extracts fiber width edges, and improves measurement accuracy. By analyzing the changes in pixel intensity near the edge and using interpolation or fitting methods, the exact position of the edge can be more accurately estimated, thereby capturing the subtle boundaries of the fiber material, ensuring that the measurement results are more reliable and accurate, and thus optimizing the detection effect of the fiber width; S23, converting pixels to actual physical sizes through pixel equivalent calibration, calculating width data, and further detecting width fluctuations.

3. The method according to claim 2, characterized in that Step S21 of roughly positioning the fiber bundle specifically includes the following method steps: performing rough positioning in sequence through distortion correction, image filtering, image enhancement, traditional edge detection operators and morphological processing.

4. The method according to claim 1, characterized in that: The target detection algorithm used in step S3 is YOLO V8.

5. The method according to claim 4, characterized in that The step of using the target detection algorithm model to perform detection in step S3 specifically includes: S31, data augmentation is performed to expand the data set, including applying a series of transformations such as rotation, flipping, scaling, cropping, color change, etc. to the original data, thereby effectively expanding the scale of the training data; S32, dividing the enhanced data into a training set and a test set, and manually annotating them; S33, modify hyperparameters and perform iterative training to optimize model performance; S34, select the optimal training model and evaluate it on the test set; S35 uses a target detection algorithm model to automatically identify and classify multiple defects using a deep learning algorithm for high-definition image acquisition and real-time analysis under high-speed operation and complex motion trajectories of the tow.

6. A computer-readable storage medium, characterized in that: A computer program is stored on the medium, and after the computer program is run, the method according to any one of claims 1 to 5 is executed.

7. A computer system, characterized in that: The method comprises a processor and a storage medium, wherein a computer program is stored in the storage medium, and the processor reads and runs the computer program from the storage medium to execute the method as claimed in any one of claims 1 to 5.

8. A system based on the method according to any one of claims 1 to 5, characterized in that: Including image acquisition module, lighting module, ribbon width detection module, and other defect detection modules; The image acquisition module is used to acquire high-resolution images of the moving fiber bundles under controlled lighting conditions, and convert the acquired image signals into digital signals, which are uploaded to the computer for further analysis by the ribbon width detection module and other defect detection modules; The lighting module is used to highlight the detection features of the object to be tested, and adopts a white strip LED light source that matches the shape of the object to be tested, so as to ensure uniform lighting and reduce the impact of ambient light on image quality, and ensure sufficient contrast between the image background and the object to be tested; The ribbon width detection module is used to calculate width data and further detect width fluctuations; The other defect detection module uses the YOLO V8 target detection algorithm to perform high-definition image acquisition and real-time analysis to automatically identify and classify various defects.

Images