A damage identification method for laser damage test points of an n×n two-dimensional array

Abstract

The present invention discloses a method for identifying damage to laser damage test points in an n×n two-dimensional array. This method locates the central position of each test point, segments the damage image of the array points according to the interval of the test dot matrix, performs relevant operations on the images before and after damage through wavelet transform and image segmentation to obtain a correlation coefficient, determines the threshold of the correlation coefficient by combining statistics and phase contrast microscopy, and realizes the rapid identification of the damage condition of the two-dimensional array test points by comparing the correlation coefficient with the threshold of the correlation coefficient, significantly improving the efficiency of damage identification.

Description

Technical Field

[0001] The present invention relates to a method for identifying damage to laser damage test points in an n×n two-dimensional array, belonging to the field of laser technology. Background Art

[0002] Laser thin films are the most vulnerable part in lasers and their application systems, and their laser damage resistance ability is the key to restricting the stability, reliability and service life of lasers and their application systems. For a long time, China's laser thin film products have been unable to break through the technical bottleneck of high laser damage resistance ability, resulting in the dependence on imported high-performance laser thin film products, restricting the development of China in high-power, high-energy lasers and their applications. The preparation of high-damage-threshold laser thin films has become a bottleneck technology restricting the development of the laser industry. Scientific evaluation and accurate measurement of the laser damage threshold of optical thin films are the prerequisites for researching and preparing high-performance optical thin films. China and the ISO organization have formulated a series of measurement standards such as GJB1487-1992, GB / T16601-2017, GB / T37275-2018 and ISO11254 for the accurate measurement of the laser damage threshold. The existing measurement methods mainly obtain a test point with a single-shot laser pulse, with a long measurement process cycle, low measurement accuracy and low efficiency. The patent

A rapid measurement device for the laser damage threshold of optical thin films, patent number: ZL201910121286.1

[0003] The purpose of the present invention is to provide a method for identifying damage to laser damage test points in an n×n two-dimensional array. This method locates the center position of each test point, divides the damaged image of the array points according to the interval of the test points, performs relevant operations on the images before and after damage after wavelet transform and image segmentation, and achieves the purpose of quickly identifying the damaged points of the array by determining the threshold of the correlation coefficient.

[0004] To achieve the above purpose, as shown in the attached Figure 1 figure, the technical solution adopted by the present invention is:

[0005] A method for identifying damage to laser damage test points in an n×n two-dimensional array is mainly divided into three steps:

[0006] The first step, segmentation of the array damage test image:

[0007] Step 1-1, obtain an image of complete damage to the laser damage test points in an n×n two-dimensional array;

[0008] Step 1-2, perform binary processing on the image using the OTSU threshold (maximum inter-class variance method);

[0009] The described OTSU threshold method is as follows: Denote t as the segmentation threshold between the foreground and the background. The proportion of foreground points in the image is p1, and the average gray value is m1. The proportion of background points in the image is p2, and the average gray value is m2. Then the total average gray value of the image is:

[0010] m = p1m1 + p2m2 (1)

[0011] The variance of the foreground and background images:

[0012] σ 2 = p1(m1 - m) 2 + p2(m2 - m) 2 (2)

[0013] When the difference between the foreground and the background is the largest, that is, the variance σ 2 is the largest, then the gray value at this time is the optimal threshold;

[0014] Step 1-3: Divide all the damaged points in the binary image processed in Step 1-2 into n×n damaged point sets according to the principle of proximity;

[0015] Step 1-4: Use the least squares method to perform an envelope circle fitting on each damaged point set to obtain the center position O i (x i , y i ) of this damaged point set;

[0016] Step 1-5: Calculate the distance L i between any two center points according to formula (3), record the maximum value L i and its corresponding two-point coordinates O M (x i , y i ) and O i (x j , y j ) and O j ;

[0017]

[0018] Step 1-6: Respectively take O i (x i , y i ) and O j (x j , y j ) as the centers of the circles, and draw circles with L M / [2(n - 1)] as the radius. Then these two circles and O i (x i , y i ) and O j (x j , yj ) The extension lines of the point connections intersect at two points P i and P j , then the length of the line segment P i P j is nL M / (n - 1);

[0019] Step 1 - 7: By translation, evenly divide the square image area with P i P j as the diagonal into n×n square sub - unit images, then the side length of the square sub - unit image is i.e., the interval of the array damage test points;

[0020] Second step: Obtain a non - damaged square image, and the size of this image is the same as the side length of the square sub - unit image in Step 1 - 7;

[0021] Third step: Identify the damage condition of the array damage test point image:

[0022] Step 3 - 1: Obtain an image irradiated by an n×n two - dimensional array laser, and the interval of the array damage test points is

[0023] Step 3 - 2: Divide the image obtained in Step 3 - 1 into n×n square image sub - units in the same way as the first step;

[0024] Step 3 - 3: Perform wavelet transform processing on the two square image sub - units obtained in the second step and Step 3 - 2 respectively, and perform wavelet denoising under two - layer wavelet decomposition;

[0025] Step 3 - 4: Calculate the correlation coefficient between each square sub - unit image processed in Step 3 - 3 and the square sub - unit image obtained in the second step according to formula (4);

[0026]

[0027] In the formula, i(x, y) and f(x, y) respectively represent the pre - damage and post - damage images after wavelet transform;

[0028] Step 3 - 5: Determine the threshold H of the correlation coefficient by combining statistics and phase - contrast microscopy T ;

[0029] Step 3 - 6: If the correlation coefficient H calculated in Step 3 - 4 ≥ H T , then the square sub - unit image is damaged, that is, the image of this damage test point is damaged.

[0030] Beneficial effects: The damage identification method for an n×n two-dimensional array laser damage test point provided by the present invention can quickly identify the damage conditions of the two-dimensional array test points, significantly improving the efficiency of damage identification. Description of the Drawings

[0031] Figure 1 is a flowchart of a damage identification method for an n×n two-dimensional array laser damage test point. Detailed Embodiments

[0032] Embodiment 1 A damage identification method for an n×n two-dimensional array laser damage test point.

[0033] The damage identification method for an n×n two-dimensional array laser damage test point of the present invention, in combination with the attached Figure 1 , is carried out according to the following steps:

[0034] First step, segmentation of the array damage test image:

[0035] Step 1-1, obtain an image of a completely damaged n×n two-dimensional array laser damage test point;

[0036] Step 1-2, perform binarization processing on the image using the OTSU threshold (maximum between-class variance method),

[0037] The OTSU threshold method is: Denote t as the segmentation threshold between the foreground and the background, the proportion of foreground points in the image is p1, the average gray level is m1, the proportion of background points in the image is p2, and the average gray level is m2. Then the total average gray level of the image is:

[0038] m = p1m1 + p2m2 (1)

[0039] Variance of the foreground and background images:

[0040] σ 2 = p1(m1 - m) 2 + p2(m2 - m) 2 (2)

[0041] When the difference between the foreground and the background is the largest, that is, the variance σ 2 is the largest, then the gray level at this time is the optimal threshold;

[0042] Step 1-3, divide all damaged points in the image binarized in Step 1-2 into n×n damaged point sets according to the principle of proximity;

[0043] Step 1-4, use the least squares method to perform envelope circle fitting on each damaged point set to obtain the center position O i (x i , y i ) of the damaged point set;

[0044] Step 1-5: Calculate the distance L between any two center points according to formula (3) i , and record L i 's maximum value L M and its corresponding two-point coordinates O i (x i , y i ); j (x j , y j );

[0045]

[0046] Step 1-6: Respectively, with O i (x i , y i ) and O j (x j , y j ) as the centers of circles, draw circles with a radius of L M / [2(n - 1)]. Then these two circles intersect the extension lines of the connections between O i (x i , y i ) and O j (x j , y j ) at two points P i and P j . Then the length of the line segment P i P j is nL M / (n - 1);

[0047] Step 1-7: Uniformly divide the square image area with P i P j as the diagonal into n×n square sub-unit images by translation. Then the side length of the square sub-unit image is which is the interval between the array damage test points;

[0048] Second step: Obtain a square image without damage, and the size of this image is the same as the side length of the square sub-unit image in Step 1-7;

[0049] Third step: Identify the damage situation of the array damage test point image:

[0050] Step 3-1: Obtain an image irradiated by an n×n two-dimensional array laser, and the interval between the array damage test points is

[0051] Step 3-2: Divide the image obtained in Step 3-1 into n×n square image sub-units in the same way as in the first step;

[0052] Step 3-3: Perform wavelet transform processing on the two square image sub-units obtained in the second step and Step 3-2 respectively, and perform two-layer wavelet decomposition under the sym6 wavelet to obtain the high-frequency part and low-frequency part of the image. Use the VisuShrink threshold method to denoise the high-frequency part, and obtain the processed image through wavelet reconstruction;

[0053] Step 3-4: Calculate the correlation coefficient between each square sub-unit image processed in Step 3-3 and the square sub-unit image obtained in the second step according to formula (4);

[0054]

[0055] In the formula, i(x, y) and f(x, y) respectively represent the pre-damage and post-damage images after wavelet transform;

[0056] Step 3-5: Determine the threshold H of the correlation coefficient by combining statistics and phase contrast microscopy T ;

[0057] Step 3-6: If the correlation coefficient H calculated in Step 3-4 is ≥H T , then the square sub-unit image is damaged, that is, the image of the damage test point is damaged.

[0058] Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention, and are not intended to limit them; 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 for some of the technical features; and these modifications or replacements do not make the essence of the corresponding technical solutions deviate from the spirit and scope of the technical solutions of the various embodiments of the present invention.

Claims

1. A damage identification method for laser damage test points of an n×n two-dimensional array, characterized in that The method comprises the following steps: First step, segmentation of the array damage test image: Step 1-1: Obtain an image of complete damage to the laser damage test points of an n×n two-dimensional array; Step 1-2: Perform binarization processing on the image of complete damage to the laser damage test points of the n×n two-dimensional array using the OTSU threshold method; The OTSU threshold method is as follows: Denote t as the segmentation threshold between the foreground and the background, the proportion of foreground points in the image as p1, the average gray value as m1, the proportion of background points in the image as p2, and the average gray value as m2. Then the total average gray value of the image is: m = p1m1 + p2m2 (1) Variance of the foreground and background images: σ 2 = p1(m1 - m) 2 + p2(m2 - m) 2 (2) When the difference between the foreground and the background is the largest, i.e., the variance σ 2 is the largest, the gray level at this time is the optimal threshold value; Step 1-3: Divide all the damaged points in the image after binarization processing in Step 1-2 into n×n damaged point sets according to the principle of proximity; Steps 1-4: Use the least squares method to perform circumcircle fitting on each set of damage points, so as to obtain the center position O of the set of damage points i (x i ,y i ); Step 1-5: Calculate the distance L between any two center points according to formula (3) i , record L i 's maximum value L M and its corresponding two-point coordinates O i (x i , y i ), O j (x j , y j ); Steps 1-6: Respectively with O i (x i , y i ) and O j (x j , y j ) as the centers, and with L M / [2(n - 1)] as the radius to draw circles. Then these two circles intersect the extension lines of the connections between O i (x i , y i ) and O j (x j , y j ) at two points P i and P j . Then the length of the line segment P i P j is n×L M / (n - 1); Steps 1-7, by means of translation, uniformly divide the square image area with P i P j as the diagonal into n×n square sub-unit images, and the side length of the square sub-unit image is i.e., the interval of the array damage test points; Second step, obtain a square image without damage, and the size of this square image without damage is the same as the side length of the square sub-unit image in Step 1-7; Third step, identify the damage condition of the array damage test point image: Step 3-1: Obtain an image irradiated by a laser in an n×n two-dimensional array, with the interval between the array damage test points being Step 3-2: Segment the image obtained in Step 3-1 into n×n square sub-unit images in the same way as in the first step; Step 3-3: Perform wavelet transform processing on the two square sub-unit images obtained in the second step and Step 3-2 respectively, and perform denoising processing on the wavelets under two-layer wavelet decomposition; Step 3-4: Calculate the correlation coefficient between each square sub-unit image after being processed in Step 3-3 and the square sub-unit image obtained in the second step according to formula (4); In the formula, i(x, y) and f(x, y) respectively represent the pre-damage and post-damage images after wavelet transform; Step 3-5: Determine the threshold H of the correlation coefficient by combining statistics and phase contrast microscopy T ; Step 3-6, the correlation coefficient H calculated in Step 3-4 ≥ H T , then the image of the square sub-unit is damaged, that is, the image of the damage test point is damaged.

Images