Polyethylene fiber quality detection method based on data fitting

Abstract

The invention provides a polyethylene fiber quality detection method based on data fitting, and relates to the field of image recognition. The method specifically comprises the following steps: 1) acquiring an original polyethylene fiber image by adopting a high-performance industrial camera; 2) performing binarization preprocessing operation on the original image to solve the problem of uneven illumination and improve the image contrast; 3) by comparing the accumulated column pixel values, the segmentation of the single-strand fiber yarn is realized; and 4) calculating the sum of the number of white pixel points in the horizontal direction for the single-strand fiber yarn image to obtain a pixel projection drawing in the horizontal direction. And 5) performing linear fitting on the horizontal projection data. And 6) calculating the fitting degree between the fitting model and the observation data through a mean square error index. And 7) setting a threshold value so as to judge whether the cellosilk has a quality problem or not. The detection method provided by the invention is simple and efficient, can effectively detect abnormal detection, is high in speed, high in precision and good in reliability, and has important significance and value for comprehensive automation of production activities.

Description

Technical Field

[0001] The present invention relates to the technical field of image recognition, and in particular to a method for detecting the quality of polyethylene fibers based on data fitting. Background Art

[0002] Polyethylene fiber (PE fiber) is a new type of material, and together with carbon fiber and aramid, it is called the world's three major high-performance fibers today. PE fibers have special properties such as high strength modulus, low density, good weather resistance, and resistance to acids, alkalis, and salts, and are widely used in fields such as marine, military, aerospace, safety protection, medical, construction, and communication. During the production process of polyethylene fibers, due to reasons such as process and equipment, defects such as stiff filaments, hair filaments, and broken filaments often appear on the fiber surface, affecting product quality. Currently, the conventional method for detecting the quality of fiber filaments is to conduct real-time visual inspection on the winding process of the finished filaments by recording personnel. However, this method easily makes the staff exhausted, and has disadvantages such as low detection efficiency, low accuracy and precision, high missed detection rate and false detection rate, greatly reducing the production efficiency of PE fibers, being unable to establish an effective PE fiber defect data management mechanism, and being difficult to meet the needs of modern production development. Summary of the Invention

[0003] In view of this, the purpose of the present invention is to provide a method for detecting the quality of polyethylene fibers based on data fitting, which can simply and efficiently identify fiber filaments with quality problems through image acquisition, preprocessing, and data analysis methods, and solve the problems of low efficiency and high cost of the currently commonly used manual visual quality detection method.

[0004] The technical solution adopted by the present invention to achieve the above purpose is as follows:

[0005] A method for detecting the quality of polyethylene fibers based on data fitting, comprising the following steps:

[0006] 1) Collect fiber filament images through an industrial camera;

[0007] 2) Perform binarization processing on the collected fiber filament images;

[0008] 3) Segment single fiber filaments from the binarized images;

[0009] 4) Perform horizontal projection on the segmented images;

[0010] 5) Detect the quality of the fiber filaments according to the pixel projection map in the horizontal direction.

[0011] The specific content of step 2) is as follows:

[0012] Set a threshold T, and use this threshold T to divide the data of the fiber filament image into a pixel group greater than T and a pixel group less than T. When the gray value of a certain point in the image is greater than T, set the pixel value of this point to the maximum value of 255. When the gray value of a certain point in the image is less than T, set the pixel value of this point to the minimum value of 0.

[0013] Take one-third of the difference between the maximum gray value of the image and the minimum gray value, and subtract it from the maximum gray value of the image to obtain the threshold T.

[0014] The specific content of step 3) is as follows:

[0015] Scan the number of white pixels in each column of the image column by column from left to right. When the number of scanned white pixels is more than the set estimated value, record the column number of this column, and use this column number as the left boundary of a single fiber filament, and continue scanning until the number of white pixels in a certain column is less than the estimated value, then stop scanning, and record the column number of this column, and use this column number as the right boundary of a single fiber filament. Crop the original image according to the coordinates determined by the left and right boundaries to obtain a single fiber filament image.

[0016] The specific content of step 4) is as follows:

[0017] Calculate the sum of the number of white pixels in the horizontal direction of each fiber filament image to obtain a pixel projection map in the horizontal direction.

[0018] Step 5) includes the following steps:

[0019] 5.1) Perform linear fitting on the pixel projection data in the horizontal direction according to the linear regression model;

[0020] 5.2) Calculate the regression coefficients by the least squares method to obtain the model prediction values, and draw the fitting curve;

[0021] 5.3) Use the mean square error as the fitting degree index to calculate the squared difference value between the model prediction value and the true value;

[0022] 5.4) If the calculated fitting degree index is less than the threshold, it is determined that the quality of the fiber filament is good.

[0023] A polyethylene fiber quality detection system based on data fitting, comprising:

[0024] An image acquisition module, used to acquire fiber filament images through an industrial camera;

[0025] A binarization processing module, used to perform binarization processing on the acquired fiber filament images;

[0026] An image segmentation module, used to segment single fiber filaments from the binarized image;

[0027] A projection module for performing horizontal projection on the segmented image;

[0028] A quality inspection module for inspecting the quality of fiber filaments based on the pixel projection map in the horizontal direction.

[0029] A polyethylene fiber quality inspection device based on data fitting, comprising a memory and a processor; the memory is used for storing a computer program; the processor is used for implementing the polyethylene fiber quality inspection method based on data fitting when executing the computer program.

[0030] A computer-readable storage medium, on which a computer program is stored, and when the computer program is executed by a processor, the polyethylene fiber quality inspection method based on data fitting is implemented.

[0031] The present invention has the following beneficial effects and advantages:

[0032] 1. The present invention can provide a method for inspecting the quality of fiber filaments;

[0033] 2. The present invention can achieve the segmentation of single-strand fiber filaments;

[0034] 3. The algorithm complexity of the present invention is low and can meet the requirements of on-line production inspection;

[0035] 4. By recording the quality inspection results, the present invention can improve the traceability of fiber quality information. Description of the Drawings

[0036] Figure 1 Example diagram of polyethylene fiber image acquisition;

[0037] Figure 2 Flowchart of a polyethylene fiber detection method based on data fitting;

[0038] Figure 3 Effect diagram of image binarization;

[0039] Figure 4 Example diagram of single-strand fiber filament image

[0040] Figure 5 Horizontal direction pixel projection map and data fitting result map. Detailed Embodiment

[0041] To make the implementation purpose, technical solution and advantages of the present invention clearer, the present invention will be further described in detail below with reference to the drawings and embodiments.

[0042] As Figure 2 shown, it specifically includes the following steps:

[0043] Image acquisition: A high-performance industrial camera is used for Gigabit Ethernet transmission, and a customized LED light source is adopted, which is stable, reliable, and has low light decay. The acquired images are stable, clear, and reliable, and on-site maintenance is simple and convenient. The acquisition frame rate can reach 13 frames per second. The customized LED light source illuminates the fiber filaments, and a high-performance industrial camera is used to acquire clear PE fiber images, including fiber filaments without quality problems and fiber filaments with different quality problems such as stiff filaments, hair filaments, and broken filaments. An example of the acquired image is shown in Figure 1 as shown.

[0044] Image binarization: A threshold T is set, and the data of the image is divided into two parts by T: the pixel group greater than T and the pixel group less than T. If the gray value of a certain point in the image is greater than this value, the point is set to the maximum value of 255, and conversely, if the gray value of a certain point in the image is less than this value, the point value is set to the minimum value of 0.

[0045] Threshold adjustment: In the present invention, for the images acquired under all-day and all-time different lighting conditions in the on-site environment, the binarization threshold is adjusted multiple times. Finally, a satisfactory threshold can be obtained by subtracting one-third of the difference between the maximum gray value and the minimum gray value of the image from the maximum gray value of the image, and the binarization effect is better. The image binarization effect is as shown in Figure 3 as shown.

[0046] Single-strand fiber filament segmentation: The number of white pixel points in each column of the cumulative image is counted, and the image is scanned column by column from left to right. When the number of white pixel points scanned is more than a certain value, the column number of this column is recorded, and this column number is the left boundary of the single-strand fiber filament. Continue scanning until the number of white pixel points in a certain column is less than the estimated value, stop scanning, and record the column number of this column, which is the right boundary of the single-strand fiber filament. Then, the original image is cropped according to the coordinates determined by the above method, and the single-strand fiber filament image can be obtained.

[0047] It should be noted that with the change of winter temperature, the fiber filaments will show different degrees of looseness due to static electricity. Therefore, according to the production equipment and process of the factory, the shortest actual distance between adjacent two strands of fiber filaments is obtained, and a threshold is set according to the column pixel value distribution. If it is greater than this threshold, it is determined as two strands of fiber filaments, and thus the independent segmentation of each strand of fiber filament in the entire image is realized.

[0048] Horizontal projection: Calculate the sum of the number of white pixel points in the horizontal direction of each strand of fiber filament image to obtain the pixel projection map in the horizontal direction.

[0049] Feature analysis: Figure 4 For the single-strand fiber filament image, it can be seen that Figure 4 (a) is a fiber filament with good quality, Figure 4 (b - d) are fiber filaments with quality problems. Figure 5 The blue curve in is for Figure 4One-to-one horizontal pixel projection diagram, as shown in Figure 5 (a), it can be seen that the pixel distribution of standard fiber filaments with good quality is smoother. Figure 5 (b-d) For fiber filaments with quality problems such as dead filaments and hair filaments, the pixel distribution is more likely to fluctuate.

[0050] Linear fitting: Based on the actual analysis of the fiber image features, prepare the horizontal pixel projection data of a single fiber filament; use a linear regression model to perform linear fitting on the horizontal projection data, calculate the regression coefficients by the least squares method, obtain the model prediction values, and draw the fitting curve ( Figure 5 the red straight line in).

[0051] Calculation of the goodness of fit: Mean Square Error (MSE) is a commonly used method for calculating the goodness of fit. It calculates the squared difference between the model prediction values and the true values. The smaller the MSE, the better the fitting degree between the model and the observed data. The advantage of MSE is that it squares the difference values, so larger error values have a greater impact on the goodness of fit, which helps to more sensitively capture the prediction errors of the model. Use the Mean Square Error (MSE) as the goodness-of-fit index to calculate the squared difference between the model prediction values and the true values, where the true values are the horizontal projection data of the single fiber filament image. Finally, determine whether there are quality problems with the fiber filaments by setting a goodness-of-fit threshold. The calculation formula is as follows:

[0052]

[0053] where, y i represents the true value, represents the predicted value, n represents the number of samples, and ∑ represents the summation operation. The index values in MSE are as shown in Figure 5 .

[0054] Quality inspection: Through the experiments in this paper, the goodness-of-fit MSE threshold is set to 100000. If it is less than this threshold, it is judged that the quality of the fiber filament is good.

Claims

1. A method for detecting the quality of polyethylene fibers based on data fitting, characterized in that, It includes the following steps: 1) Collect fiber filament images through an industrial camera; 2) Perform binarization processing on the collected fiber filament images; 3) Segment single fiber filaments from the binarized images; 4) Perform horizontal projection on the segmented images; 5) Conduct quality inspection on the fiber filaments according to the pixel projection map in the horizontal direction.

2. The method for detecting the quality of polyethylene fibers based on data fitting according to claim 1, characterized in that, The specific content of step 2) is as follows: Set a threshold T, and use this threshold T to divide the data of the fiber filament image into a pixel group greater than T and a pixel group less than T. When the gray value of a certain point in the image is greater than T, set the pixel value of this point to the maximum 255. When the gray value of a certain point in the image is less than T, set the pixel value of this point to the minimum 0.

3. The method for detecting the quality of polyethylene fibers based on data fitting according to claim 2, characterized in that, Take one-third of the difference between the maximum gray value of the image and the minimum gray value and subtract it from the maximum gray value of the image as the threshold T.

4. The method for detecting the quality of polyethylene fibers based on data fitting according to claim 1, characterized in that, The specific content of step 3) is as follows: Scan the number of white pixel points in each column of the image column by column from left to right. When the number of white pixel points scanned is more than the set estimated value, record the column number of this column, and use this column number as the left boundary of the single fiber filament, and continue scanning until the number of white pixel points in a certain column is less than the estimated value, then stop scanning, and record the column number of this column, and use this column number as the right boundary of the single fiber filament. Crop the original image according to the coordinates determined by the left and right boundaries to obtain a single fiber filament image.

5. The method for detecting the quality of polyethylene fibers based on data fitting according to claim 1, characterized in that, The specific content of step 4) is as follows: Calculate the sum of the number of white pixel points in the horizontal direction of each fiber filament image to obtain a pixel projection map in the horizontal direction.

6. The method for detecting the quality of polyethylene fibers based on data fitting according to claim 1, characterized in that, The specific content of step 5) includes the following steps: 5.1) Perform linear fitting on the pixel projection data in the horizontal direction according to the linear regression model; 5.2) Calculate the regression coefficient through the least squares method to obtain the model prediction value, and draw a fitting curve; 5.3) Use the mean square error as the fitting degree index to calculate the squared difference value between the model prediction value and the true value; 5.4) If the calculated fitting degree index is less than the threshold, it is determined that the quality of the fiber filament is good.

7. A system for detecting the quality of polyethylene fibers based on data fitting, characterized in that, It includes: An image acquisition module, which is used to collect fiber filament images through an industrial camera; A binarization processing module, which is used to perform binarization processing on the collected fiber filament images; An image segmentation module, which is used to segment single fiber filaments from the binarized images; A projection module, which is used to perform horizontal projection on the segmented images; A quality inspection module, which is used to conduct quality inspection on the fiber filaments according to the pixel projection map in the horizontal direction.

8. A device for detecting the quality of polyethylene fibers based on data fitting, characterized in that, It includes a memory and a processor; the memory is used to store computer programs; the processor is used to, when executing the computer programs, implement a polyethylene fiber quality inspection method based on data fitting as described in any one of claims 1-6.

9. A computer-readable storage medium, characterized in that, A computer program is stored on the storage medium, and when the computer program is executed by the processor, a polyethylene fiber quality inspection method based on data fitting as described in any one of claims 1-6 is implemented.

Images