CN-120047407-B - Defect identification method and system for molybdenum-rhenium alloy pipe fitting based on high-definition microscopic image

CN120047407BCN 120047407 BCN120047407 BCN 120047407BCN-120047407-B

Abstract

The invention discloses a defect identification method and a defect identification system for a molybdenum-rhenium alloy pipe fitting based on high-definition microscopic images, wherein the method comprises the following operation steps of acquiring high-definition images of the molybdenum-rhenium alloy pipe fitting by utilizing high-definition microscopic equipment; the method comprises the steps of preprocessing a high-definition image of a molybdenum-rhenium alloy pipe fitting to obtain a molybdenum-rhenium image to be detected, performing execution processing of a region growing algorithm on the molybdenum-rhenium image to be detected to obtain a defect region to be detected, identifying defects in the defect region to be detected to obtain type defects, and judging that the molybdenum-rhenium alloy pipe fitting has defect problems through the type defects.

Inventors

LI BOBO
ZHANG XUEFENG
GUO YAJUN
NING LAIYUAN
MAO BINGLONG
LV SHANGJIE
SHI YIBO
HUANG JUNJIE

Assignees

丰联科光电(洛阳)股份有限公司

Dates

Publication Date: 20260505
Application Date: 20250124

Claims (7)

1. The defect identification method of the molybdenum-rhenium alloy pipe fitting based on the high-definition microscopic image is characterized by comprising the following operation steps of: preprocessing the high-definition image of the molybdenum-rhenium alloy pipe fitting to obtain a molybdenum-rhenium image to be detected; Performing execution processing of an area growing algorithm on the molybdenum-rhenium image to be detected to obtain a defect area to be detected; Judging that the molybdenum-rhenium alloy pipe fitting has a defect problem through the type defect; the type defects comprise air hole type defects and crack type defects; Performing an area growth algorithm on the molybdenum-rhenium image to be detected to obtain a defect area to be detected, wherein the specific operation steps are as follows: Calculating gray values of each pixel point in the molybdenum rhenium image to be detected by using an Otsu algorithm, counting pixel frequencies by the gray values of the pixel points, and carrying out normalization processing on the pixel frequencies to obtain pixel probability; constructing a gray level histogram of the molybdenum rhenium image to be detected through pixel frequency; Calculating an inter-class variance through the gray level histogram by carrying out separation degree between a foreground and a background on a molybdenum-rhenium image to be detected, wherein the calculation formula of the inter-class variance of the separation degree is as follows: ; wherein t is a threshold value, expressed as a gray scale of 0 to 255; Represented as a sum of pixel probabilities below a threshold t; represented as a sum of pixel probabilities above a threshold t; expressed as an average gray value below a threshold t; average gray values above the threshold t; traversing all the gray scales from 0 to 255 of the threshold t, and calculating the inter-class variance corresponding to the separation degree Obtaining a standard threshold T, wherein the calculation formula is as follows: ; setting a low threshold according to the standard threshold T And a high threshold ; According to the low threshold value Standard threshold T and high threshold Constructing a multi-level threshold sequence; Calculating the ambiguity index of each pixel point in the molybdenum-rhenium image to be detected, and constructing a fuzzy entropy matrix, extracting a characteristic diagram of the pixel point through the fuzzy entropy matrix, and carrying out weighted scoring on each pixel point in the characteristic of the pixel point to obtain the confidence score of each pixel point; Screening strong pixel points and weak pixel points for each pixel point of the confidence coefficient score by constructing a multi-level threshold sequence to obtain seed points; the seed points are divided into strong seed points and weak seed points; The method comprises the steps of growing a neighborhood weak pixel point by using a region growing algorithm, and calculating Euclidean distance between the weak seed point and the neighborhood weak pixel point; judging whether the Euclidean distance between the weak seed point and the neighborhood weak pixel point is smaller than a preset Euclidean distance threshold h or not; if not, the weak seed point does not meet the growth condition, and stopping growth; If yes, the weak seed points grow the neighborhood weak pixel points until the weak seed points do not meet the growth conditions, the growth is stopped, and finally a defect area to be detected is obtained and is used as the weak defect area to be detected; Growing the neighborhood strong pixel points on the strong seed points; calculating Euclidean distance between the strong seed point and the neighborhood strong pixel point; Judging whether the Euclidean distance between the strong seed point and the neighborhood strong pixel point is smaller than a preset Euclidean distance threshold h or not; If not, the strong seed point does not meet the growth condition, and stopping growth; If yes, the strong seed point grows the neighborhood weak pixel point until the strong seed point does not meet the growth condition, and the growth is stopped, and finally, the defect area to be detected is obtained and is used as the strong defect area to be detected.
2. The defect identification method of the molybdenum-rhenium alloy pipe fitting based on the high-definition microscopic image according to claim 1, wherein the defect identification method is characterized in that defects are identified in the defect area to be detected, and type defects are obtained, and the specific operation steps are as follows: Clustering the same pixel gray values of the pixels in the weak defect area to be detected to obtain a plurality of first pixel clustering clusters; Carrying out morphological pretreatment on the weak to-be-detected defect area, and enhancing the boundary contour of the suspected pore contour in the weak to-be-detected defect area; Calculating the area and perimeter of the suspected pore contour according to the boundary contour of the suspected pore contour; analyzing the pore circularity of the suspected pore profile according to the area and perimeter of the suspected pore profile; Presetting an air hole roundness threshold o, and judging whether the air hole roundness of the suspected air hole outline is larger than the air hole roundness threshold o; if not, the outline of the suspected air hole is in a regular shape, and the defect area to be detected weakly is judged to have no air hole type defect; if yes, judging that the outline of the suspected air hole presents an irregular shape, and judging that the defect area to be detected weakly has an air hole type defect; Selecting a pixel point with the maximum gray value of the pixel point from the strong defect area to be detected as a starting pixel point, and judging whether the gray value of the starting pixel point is smaller than or equal to a neighborhood pixel point by using an edge detection algorithm; if not, stopping growing the initial pixel point, wherein the strong defect area to be detected has no crack type defect; If yes, the initial pixel point continues to grow until the initial pixel point grows to form a crack contour of a strong defect area to be detected, and judging that the strong defect area to be detected has crack type defects; And if the growth of the initial pixel point is stopped during the growth process, judging that the strong defect area to be detected has no crack type defect.
3. The defect identification method for the molybdenum-rhenium alloy pipe fitting based on the high-definition microscopic image according to claim 2 is characterized by comprising the following steps of calculating a ambiguity index of each pixel point in a molybdenum-rhenium image to be detected, constructing a fuzzy entropy matrix, extracting a characteristic diagram of the pixel point through the fuzzy entropy matrix, and specifically operating the following steps: Setting a multiplication proportion coefficient, and decomposing the molybdenum-rhenium image to be detected by using a Gaussian pyramid according to the multiplication proportion coefficient to obtain a plurality of scale molybdenum-rhenium images; performing window smoothing of Gaussian kernels with different sizes on the multiple scale molybdenum-rhenium images, and continuously performing downsampling by bilinear interpolation to obtain downsampled multiple scale molybdenum-rhenium images; setting a pixel neighborhood window, and calculating local mean and local standard deviation of the neighborhood pixel according to the pixel neighborhood window for each pixel in the molybdenum-rhenium images with multiple scales; Calculating the difference between the gray value of the pixel point of each pixel point and the local mean value, and calculating the normalized ambiguity index of the difference to generate an ambiguity mapping matrix of each scale molybdenum-rhenium image; Acquiring a fuzzy entropy feature map through a fuzzy index of each pixel point of the fuzzy entropy matrix; traversing the variance of the neighborhood pixel points of each pixel point to obtain the gray dispersion degree of the neighborhood pixel points of each pixel point, and obtaining a gray feature map of each pixel point according to the gray dispersion degree of the neighborhood pixel points of each pixel point; determining the maximum pixel gray value and the minimum pixel gray value of the pixel neighborhood in the pixel neighborhood window, calculating according to the maximum pixel gray value and the minimum pixel gray value to obtain the difference contrast, and traversing to calculate the difference contrast in the pixel neighborhood window of each pixel to obtain the contrast characteristic map.
4. The defect identification method for the molybdenum-rhenium alloy pipe fitting based on the high-definition microscopic image according to claim 3, wherein the method is characterized in that each pixel point in the pixel point characteristics is weighted and scored to obtain the confidence score of each pixel point, each pixel point with the confidence score is screened by constructing a multi-level threshold sequence to obtain a seed point, and the specific operation steps are as follows: Carrying out weighted evaluation on the pixel points by the fuzzy entropy feature map, the gray feature map and the contrast feature map of the pixel points in the molybdenum rhenium image to be detected, and calculating to obtain the confidence coefficient score of the pixel points; Traversing confidence scores of each pixel point to determine whether the confidence score is greater than a low threshold in the multi-level threshold sequence ; If not, the pixel point is a weak pixel point and is used as a weak seed point; if yes, continuously judging whether the confidence score of the residual pixel points is larger than a high threshold value ; If not, the pixel point is a normal pixel point in the molybdenum rhenium image to be detected; If yes, the rest pixel points are taken as strong pixel points and are taken as strong seed points.
5. The defect identification method based on high-definition microscopic images for molybdenum-rhenium alloy pipe fittings according to claim 4, wherein the method is characterized in that the fuzzy entropy feature map, gray feature map and contrast feature map of the pixel points in the molybdenum-rhenium images to be detected are used for carrying out weighted evaluation on the pixel points, and the confidence score of the pixel points is calculated, and the specific operation steps are as follows: Normalizing the fuzzy entropy feature map, the gray feature map and the contrast feature map, and combining the fuzzy entropy feature map, the gray feature map and the contrast feature map by using a linear weighting method to obtain a fusion feature; and carrying out weighted evaluation through the fusion characteristics to obtain the confidence score of the pixel point.
6. The defect identification method for the molybdenum-rhenium alloy pipe fitting based on the high-definition microscopic image according to claim 5, wherein the defect problem of the molybdenum-rhenium alloy pipe fitting is judged by the type defect, and the specific operation steps are as follows: And when any type of defects including the pore type defects or the crack type defects exist in the molybdenum-rhenium image to be detected, judging that the defect problem exists in the molybdenum-rhenium image to be detected.
7. A defect identification system of a molybdenum-rhenium alloy pipe fitting based on a high-definition microscopic image, which is used for realizing the defect identification method of the molybdenum-rhenium alloy pipe fitting based on the high-definition microscopic image according to any one of claims 1-6, and is characterized by comprising an acquisition module, an identification module, a conclusion module, a detection module and a detection module; The acquisition module is used for acquiring high-definition images of the molybdenum-rhenium alloy pipe fitting by utilizing high-definition microscopic equipment, and preprocessing the high-definition images of the molybdenum-rhenium alloy pipe fitting to obtain molybdenum-rhenium images to be detected; the identification module is used for performing execution processing of an area growth algorithm on the molybdenum-rhenium image to be detected to obtain a defect area to be detected; And the conclusion module is used for judging that the molybdenum-rhenium alloy pipe fitting has a defect problem through the type defect.

Description

Defect identification method and system for molybdenum-rhenium alloy pipe fitting based on high-definition microscopic image Technical Field The invention relates to anomaly detection, in particular to a defect identification method and system for a molybdenum-rhenium alloy pipe fitting based on high-definition microscopic images. Background Molybdenum-rhenium alloy is highly valued as an important superalloy material because of its excellent high temperature properties, good corrosion resistance and mechanical strength. As one of the main forms, molybdenum-rhenium alloy pipe fittings are often affected by temperature changes, mechanical stresses, chemical reactions and other factors during manufacture and use, and various types of defects, mainly including pores and cracks, are generated. Porosity defects are typically voids formed in or on the metal due to incomplete escape of gas during the smelting process, and cracks may result in cracking of the metal surface or interior due to excessive stretching, fatigue, or uneven cooling, etc. These defects affect not only the mechanical properties of the molybdenum-rhenium alloy tube, but also the reliability of its use in extreme environments. The traditional defect detection method, such as manual visual inspection or a method based on common image processing, often has larger risks of detection omission and misjudgment due to the problems of insufficient image resolution, tiny defects, difficult identification and the like, and is difficult to meet the requirements of high precision and high reliability. At present, an image segmentation algorithm is used as an important defect detection means and has been widely applied to various industrial detection. Particularly, the region growing algorithm can effectively extract a defective region in an image by gradually expanding a detection region from a specific seed point. However, how to improve the segmentation accuracy of the region growing algorithm and further identify the type of defects based on these segmented regions is a key issue in current research, because the surface defects of molybdenum-rhenium alloy pipe fittings are complex in morphology and may be similar to the surrounding background. Disclosure of Invention The invention aims to provide a defect identification method and system for a molybdenum-rhenium alloy pipe fitting based on high-definition microscopic images, and solves the technical problems pointed out in the prior art. The invention provides a defect identification method of a molybdenum-rhenium alloy pipe fitting based on high-definition microscopic images, which comprises the following operation steps: preprocessing the high-definition image of the molybdenum-rhenium alloy pipe fitting to obtain a molybdenum-rhenium image to be detected; Performing execution processing of an area growing algorithm on the molybdenum-rhenium image to be detected to obtain a defect area to be detected; Judging that the molybdenum-rhenium alloy pipe fitting has a defect problem through the type defect. Preferably, as one embodiment, the type defects include a void type defect and a crack type defect. Preferentially, as an implementation scheme, the method for detecting the defect area comprises the following steps of: Calculating gray values of each pixel point in the molybdenum rhenium image to be detected by using an Otsu algorithm, counting pixel frequencies by the gray values of the pixel points, and carrying out normalization processing on the pixel frequencies to obtain pixel probability; constructing a gray level histogram of the molybdenum rhenium image to be detected through pixel frequency; Calculating an inter-class variance through the gray level histogram by carrying out separation degree between a foreground and a background on a molybdenum-rhenium image to be detected, wherein the calculation formula of the inter-class variance of the separation degree is as follows: ; wherein t is a threshold value, expressed as a gray scale of 0 to 255; Represented as a sum of pixel probabilities below a threshold t; represented as a sum of pixel probabilities above a threshold t; expressed as an average gray value below a threshold t; average gray values above the threshold t; traversing all the gray scales from 0 to 255 of the threshold t, and calculating the inter-class variance corresponding to the separation degree Obtaining a standard threshold T, wherein the calculation formula is as follows: ; setting a low threshold according to the standard threshold T And a high threshold; According to the low threshold valueStandard threshold T and high thresholdConstructing a multi-level threshold sequence; Calculating the ambiguity index of each pixel point in the molybdenum-rhenium image to be detected, and constructing a fuzzy entropy matrix, extracting a characteristic diagram of the pixel point through the fuzzy entropy matrix, and carrying out weighted scoring on each pixel point in the characteristic