Test sieve calibration method based on machine vision

Abstract

The invention relates to the technical field of measurement and detection, in particular to a test sieve calibration method based on machine vision. Comprising the following steps: acquiring a test screen image by a microscope, firstly performing graying processing on an original image, and converting a color image into a grayscale image; the method comprises the following steps: carrying out binarization processing on a grey-scale image, carrying out binarization on the image, setting a grey-scale value of a pixel point on the image to be 0 or 255, enabling the whole image to present an obvious visual effect which is only black and white, and better analyzing the shape and the contour of an object through binarization; and performing connected domain analysis on the binary image, finding a pixel point to which each sieve hole belongs, endowing each pixel point with a label through the connected domain analysis, and forming a connected domain by the pixel points with the same label value so as to realize segmentation of the region of interest. The method is applied to the measurement calibration work of the test sieve, the working efficiency of verification and calibration personnel can be greatly improved by using the method, and human resources are saved.

Description

Technical Field

[0001] This invention relates to the field of metrology and testing technology, and specifically to a test sieve calibration method based on machine vision. Background Technology

[0002] Test sieves are sieves that meet the technical requirements of test sieve standards and are used for particle size analysis using sieving methods. They mainly consist of a sieve surface and a sieve frame. Based on the structure of the sieve surface, they can be divided into metal wire woven mesh test sieves, metal perforated plate test sieves, and electroformed thin plate test sieves. The national standard GB6003-1985 "Test Sieves" provides unified regulations on the requirements and marking methods for test sieves produced in my country. In many industries, particle size control is an extremely important product quality element. For example, in particle size analysis and sieving tests conducted in industries such as geological exploration, metallurgy, chemical industry, building materials, and abrasives, particle size control directly affects the final quality or survey results. In scientific research and design, the particle size composition of the raw materials processed and the products produced is important design calculation data. Therefore, test sieves, as a general-purpose measuring instrument, are of great significance to production, scientific research, and design in the national economy. Due to limitations in sieve weaving technology, the sieve apertures vary in size. During use, abrasion of the sieve wires and particle swelling can cause deformation of the sieve apertures, resulting in inaccurate sieving results. Therefore, test sieves must be calibrated frequently.

[0003] Currently, the testing and calibration of test sieves requires manual operation using a tool microscope or projector and vernier calipers. According to JJF 1175—2021 "Test Sieve Calibration Specification", all sieves with no more than 20 mesh openings must be tested, and for sieves with more than 20 mesh openings, at least 40 mesh openings must usually be tested.

[0004] Current calibration techniques require the use of a universal tool microscope or projector to measure the test sieves, and necessitate manual alignment of each mesh. Each measurement must be taken by placing the center of the crosshair on the center of the edge of the warp or weft mesh, repeating this process 20 to 40 times. Even with skilled operation and a smooth testing process, this takes 10 to 20 minutes, averaging 0.5 minutes per sieve opening. In daily testing, most test sieves have openings larger than 40mm. As testing time increases, operators inevitably experience a decline in attention and concentration, and the influence of the instrument's field of view cannot be eliminated. Therefore, traditional test sieve testing data is subjective and arbitrary, and calibration methods suffer from drawbacks such as high workload, low efficiency, and the introduction of numerous uncertainties. Summary of the Invention

[0005] Based on the above-mentioned technical problems, this invention proposes a test sieve calibration method based on machine vision to solve the problems of high randomness, heavy workload and low detection efficiency of existing test sieve calibration methods, thereby optimizing the detection method and improving detection efficiency.

[0006] To address the aforementioned technical problems, one objective of this invention is to provide a machine vision-based test sieve calibration method, the specific steps of which are as follows: S1: The universal tool microscope acquires the test sieve image. First, the original image is grayscaled using formula (1). Image grayscale is the process of converting a color image into a grayscale image. The color of each pixel is determined by three components (red, green, blue, RGB). In a grayscale image, the values ​​of these three components are the same, i.e., grayscale value. S2: Binarize the grayscale image. Image binarization is to set the grayscale value of the pixels in the image to 0 or 255, which means that the entire image presents a visual effect of only black and white. Through binarization, the shape and outline of the object can be analyzed better. S3: Perform connected component analysis on the binary image to find the pixel to which each sieve hole belongs. Connected component analysis assigns a label to each pixel. Pixels with the same label value form a connected component, thus achieving the segmentation of the region of interest.

[0007] Furthermore, S1 uses a weighted average method to convert the image to grayscale. Based on the human eye's sensitivity to different colors, different weights are assigned to perform a weighted average. Typically, green has the highest sensitivity, so the green component also has the highest weight.

[0008] Furthermore, S1 is defined by the formula: The original image is converted to grayscale.

[0009] Furthermore, S2 employs a region adaptive method to binarize the image, allowing each small region of the image to use a different threshold, thus preserving the important features of the image.

[0010] Furthermore, the specific operation of S2 is as follows: S21: Select a resolution size of The area; S22: Move this region in the image with a fixed step size; S23: Use the average value of all pixels in the region as the threshold of this region. If the pixel value of a pixel in this region is less than the threshold, set it to 0; otherwise, set it to 255. S24: After traversing all pixels of the image, complete the binarization of the image.

[0011] Furthermore, S3 uses the seed filling method to perform connected component analysis.

[0012] Furthermore, the specific steps of the seed filling method are as follows: S31: Starting from a seed, search outwards to the surrounding area. If equal pixel values ​​are found, they are marked with the same label. S32: Continue searching the neighborhood until there are no more identical pixel values ​​around it, then a connected region is found; S33: Continue searching for the next connected region using other seeds.

[0013] Furthermore, the seed is a pixel of interest (with a value greater than 1).

[0014] The above-described one or more technical solutions of the present invention have at least one or more of the following technical effects: the method used in this method has better repeatability; it greatly improves work efficiency without sacrificing detection accuracy; and it saves labor costs. Attached Figure Description

[0015] Figure 1 a: Original image of the test sieve obtained by microscope; Figure 1 b: Grayscale image of the test sieve obtained by microscope; Figure 2 Binarized image; Figure 3 Test results. Detailed Implementation

[0016] This invention proposes an automatic calibration method for test sieves based on machine vision, applicable to the metrological calibration of test sieves. This method primarily utilizes image processing algorithms, including image denoising, grayscale conversion, binarization, and connected component analysis, to calculate sieve aperture data, achieving efficient acquisition of test sieve calibration results and thus enabling automatic detection of test sieves. When applied to the metrological calibration of test sieves, this method can significantly improve the work efficiency of calibration personnel and save human resources.

[0017] The technical solutions of the present invention will be clearly and completely described below with reference to the embodiments and accompanying drawings. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments obtained. All other embodiments obtained by those skilled in the art based on the embodiments of the present invention are within the scope of protection of the present invention.

[0018] Images of test sieves obtained by a universal tool microscope, such as Figure 1 As shown in (a). First, the original image is converted to grayscale using equation (1), and the result is as follows. Figure 1As shown in (b), image grayscale conversion is the process of converting a color image to a grayscale image. In a grayscale image, the color of each pixel is determined by three components (red, green, and blue, RGB). In a grayscale image, these three components have the same value, i.e., the grayscale value. The purpose of grayscale conversion is to simplify image processing, reduce computational complexity, and preserve the image's basic information. This paper uses a weighted average method for image grayscale conversion. Based on the human eye's sensitivity to different colors, different weights are assigned for weighted averaging. Typically, green has the highest sensitivity, therefore the green component has the highest weight.

[0019] (1) Then, the grayscale image is binarized. Image binarization sets the grayscale value of each pixel in the image to 0 or 255, resulting in a visual effect where the entire image consists only of black and white. Binarization allows for better analysis of the shape and contours of objects. This paper uses a region-adaptive method for image binarization. This technique allows different thresholds to be used for each small region of the image, thus better preserving important image features such as edges and details. Compared to traditional fixed-threshold binarization, it provides better processing results, especially when there are significant uneven lighting or contrast variations in the image. First, a resolution of [resolution value missing] is selected. The image is binarized by traversing a region and then moving this region across the image with a fixed step size. The average value of all pixels within this region is used as the threshold. If a pixel value within this region is less than the threshold, it is set to 0; otherwise, it is set to 255. After traversing all pixels in the image, the image binarization is complete, and the result is as follows: Figure 2 As shown.

[0020] Next, connected component analysis is performed on the binary image to find the pixel corresponding to each sieve hole. Connected component analysis assigns a label to each pixel; pixels with the same label value form a connected component, thus achieving the segmentation of the region of interest. A connected component is a region in an image that has the same pixel value and is adjacent to each other; they are generally divided into 4-neighborhoods and 8-neighborhoods. This paper uses the seed filling method for connected component analysis. The seed filling method starts from a seed and searches outwards to the surrounding neighborhood. If equal pixel values ​​are found, they are marked with the same label, and the search continues in the neighborhood until no identical pixel values ​​are found, thus finding a connected region. Then, using other seeds, the search continues to find the next connected region. Here, the seed is a pixel of interest (with a value greater than 1).

[0021] The specific process is as follows: Iterate through the image, and if a pixel value is 1, perform the following operation: (1) Take the current pixel as a seed and assign it a new label, then add the positions of all the pixels of interest (values ​​greater than 1) in its neighborhood to the stack; (2) Pop the top pixel from the stack, assign it the same label, and add the positions of all pixels of interest (values ​​greater than 1) in its neighborhood to the stack; (3) Repeat step b until the stack is empty. At this point, a connected region is found and is marked with the same label.

[0022] After finding all connected components in the binary image, traversing all connected components and removing those with smaller areas can remove noise from the image. The remaining connected components are the ones containing the sieve holes. This paper uses the minimum bounding rectangle of each connected component as the detection result of the sieve hole. The width and height of the rectangle represent the detection results of the sieve hole in the x and y directions, respectively, expressed as pixel values. Figure 3 As shown. Finally, the standard rectangle on the universal calibration plate was used to convert the pixel values ​​to the actual dimensions, thus converting the test sieve pixel values ​​to actual values.

[0023] Obviously, those skilled in the art can make various modifications and variations to the embodiments of the present invention without departing from the spirit and scope of the embodiments of the present invention. Thus, if these modifications and variations to the embodiments of the present invention fall within the scope of the claims of the present invention and their equivalents, the present invention also intends to include these modifications and variations.

[0024] After multiple comparative experiments, the experimental results were found to be reliable, greatly improving the detection efficiency and verifying the rationality and feasibility of the proposed solution.

Claims

1. A machine vision based test screen calibration method, characterized by, The specific steps are as follows: S1: The microscope acquires a test screen image. First, the original image is subjected to grayscale processing. Image grayscale is the process of converting a color image into a grayscale image, wherein the color of each pixel is determined by three components (red, green, blue, RGB), and in a grayscale image, the values of the three components are the same, i.e., the grayscale value; S2: The grayscale image is subjected to binarization processing. Image binarization sets the grayscale values of the pixel points on the image to 0 or 255, presenting a clear visual effect of only black and white. Through binarization, the shape and contour of the object can be better analyzed; S3: The binarized image is subjected to connected component analysis to find the pixel points belonging to each screen hole. Connected component analysis assigns a label to each pixel point. Pixel points with the same label value form a connected component, achieving segmentation of the region of interest.

2. The machine vision-based test screening calibration method of claim 1, wherein: S1 uses a weighted average method for image grayscale processing. According to the sensitivity of the human eye to different colors, different weights are given for weighted average. Generally, green has the highest sensitivity, so the weight of the green component is also the highest.

3. The machine vision-based test screening calibration method of claim 2, wherein: The S1 is through the formula: The original image is grayed.

4. The machine vision-based test screening calibration method of claim 1, wherein: S2 uses a region adaptive method for image binarization. Each small region of the image is allowed to use different threshold values, preserving important features of the image.

5. The machine vision-based test screening calibration method of claim 4, wherein: The specific operation of S2 is as follows: S21: select a region with a resolution size of ; S22: Move this region in the image by a fixed step size; S23: Take the average value of all pixel values in the region as the threshold value of this region. If the pixel value of a pixel point in this region is less than the threshold value, set it to 0, otherwise set it to 255; S24: After traversing all pixel points in the image, complete the binarization of the image.

6. The machine vision-based test screening calibration method of claim 1, wherein: S3 uses a seed filling method for connected component analysis.

7. The machine vision-based test screening calibration method of claim 6, wherein, The specific steps of the seed filling method are as follows: S31: Start from a seed and search around the field. If a pixel with the same value is found, mark it as the same label; S32: Continue searching in the field until no pixel with the same value is found in the surrounding area, and then find a connected component; S33: Take another seed and continue searching for the next connected component.

8. The machine vision-based test screening calibration method of claim 7, wherein: The seed is a pixel of interest with a pixel value greater than 1.

Images