CN-122023973-A - Shaping defect detection method for silicon carbide thin-wall tube preparation process

CN122023973ACN 122023973 ACN122023973 ACN 122023973ACN-122023973-A

Abstract

The invention relates to the field of image processing, in particular to a molding defect detection method for a silicon carbide thin-wall tube preparation process. The method comprises the steps of obtaining a surface image, recording working condition identification and spatial position information, dividing the surface image into at least two local subregions, screening abnormal subregions, obtaining preliminary abnormal quantities, clustering the abnormal subregions into defect candidate regions, obtaining edge average gradient values, calculating development fluctuation complexity, calculating development fluctuation working condition response dispersion, obtaining the number of edge lines of the defect candidate regions under different working conditions, calculating development to form consistent attenuation, obtaining texture gradient main directions and development response intensities of the defect candidate regions under different working conditions, calculating development structure response consistency indexes, and generating constraint factors. The invention can improve the detection precision in the preparation process of the silicon carbide thin-wall tube.

Inventors

HUANGFU BINGCHEN
BAI XIAOXIONG
YU FEI
XING TAO
MA LIANGYUAN

Assignees

陕西固勤材料技术有限公司

Dates

Publication Date: 20260512
Application Date: 20260413

Claims (10)

1. A molding defect detection method for a silicon carbide thin-wall tube manufacturing process, comprising: Acquiring surface images of the silicon carbide thin-wall tube under different working conditions, and recording the working condition identification and the spatial position information corresponding to each image; Dividing the surface image into at least two local subregions based on the space position information, screening abnormal subregions in the surface image, acquiring preliminary abnormal quantity of each abnormal subregion, and clustering the abnormal subregions into at least one defect candidate region; performing edge detection on each defect candidate region, obtaining edge average gradient values of all abnormal subareas in each defect candidate region, and calculating the imaging fluctuation complexity of the defect candidate region based on the edge average gradient values; Acquiring the imaging fluctuation complexity of any defect candidate region under different working conditions, and calculating the response dispersion of the imaging fluctuation working condition of the defect candidate region based on the difference of the imaging fluctuation complexity under different working conditions; Acquiring the quantity of edge lines of the defect candidate region under different working conditions, and calculating the consistent attenuation of the development composition of the defect candidate region based on the development fluctuation working condition response dispersion and the change condition of the quantity of the edge lines along with the working condition; Acquiring the main direction of texture gradient and development response intensity of the defect candidate region under different working conditions, and calculating the development structure response consistency index of the defect candidate region; Generating a constraint factor based on the consistent attenuation of the imaging composition and the response consistency index of the imaging structure, modulating the preliminary abnormal quantity of the defect candidate area by using the constraint factor, and outputting a real defect detection value.
2. The method for detecting forming defects in a process of preparing a thin-wall tube of silicon carbide according to claim 1, wherein the dividing the surface image into at least two partial sub-areas based on the spatial position information, screening abnormal sub-areas therein, and obtaining preliminary abnormal amounts of each abnormal sub-area specifically comprises: Based on the space position information, carrying out gridding division on the surface image in the axial direction and the circumferential direction to obtain at least two local subareas, wherein the space position information comprises the axial coordinate and the circumferential angle information of the thin-wall tube; Extracting image characteristics of each local subarea, wherein the image characteristics comprise at least one of average gray values, gray distribution variances and texture roughness; comparing the image characteristics of each local subarea with a preset normal characteristic threshold value; if the comparison result exceeds a preset normal characteristic threshold value, judging the local subarea as an abnormal subarea; and calculating the preliminary abnormal quantity of each abnormal subarea based on the deviation degree of the image characteristic of each abnormal subarea from a preset normal characteristic threshold value.
3. The method for detecting forming defects in a process for preparing a thin-wall tube of silicon carbide according to claim 1, wherein the clustering of abnormal subregions into at least one defect candidate region specifically comprises: Establishing the spatial position indexes of all abnormal subareas; Judging whether the abnormal subareas are adjacent or not based on a preset space adjacent judging rule and a space position index, and judging that the two abnormal subareas are adjacent in space if the axial distance and the circumferential angle difference between the two abnormal subareas are smaller than a preset threshold value; merging the abnormal subareas adjacent to each other in space into the same set to form a defect candidate area; A single abnormal sub-region, which does not have any spatial adjacent relationship, is individually determined as a defect candidate region.
4. The method for detecting forming defects in a process of manufacturing a thin-wall tube of silicon carbide according to claim 1, wherein the edge detection is performed on each defect candidate region, the edge average gradient value of all abnormal sub-regions in each defect candidate region is obtained, and the complexity of the imaging fluctuation of the defect candidate region is calculated based on the edge average gradient value, specifically comprising: performing edge detection on each abnormal subarea in the defect candidate area to obtain edge pixel points; Calculating the average value of the gradient magnitudes of all the edge pixel points in each abnormal subarea, and taking the average value as the average gradient value of the edge of the abnormal subarea; arranging all abnormal subareas in the defect candidate area according to the sequence of the spatial positions of the abnormal subareas to obtain a sequencing result; Calculating absolute values of differences between edge average gradient values of adjacent abnormal subareas in each pair of sequencing results in the defect candidate areas, and calculating a first average value of the absolute values of the differences of all adjacent pairs; Calculating absolute values of differences of edge average gradient values between every two abnormal subareas in the defect candidate area, and solving a second average value of all the absolute values of the differences; Multiplying the first average value by the second average value to obtain the imaging fluctuation complexity of the defect candidate region.
5. The method for detecting the forming defect in the process of preparing the silicon carbide thin-wall tube according to claim 1, wherein the obtaining the imaging fluctuation complexity of any defect candidate region under different working conditions, and calculating the response dispersion of the imaging fluctuation working condition of the defect candidate region based on the difference of the imaging fluctuation complexity under different working conditions, specifically comprises: acquiring the imaging fluctuation complexity of the same defect candidate region under at least three different working conditions; Calculating the absolute value of the difference between the imaging fluctuation complexity corresponding to each two continuous working conditions; And (3) calculating an average value of absolute values of the display fluctuation complexity differences of all the continuous working conditions, and taking the average value as the display fluctuation working condition response dispersion of the defect candidate area.
6. The method for detecting forming defects in a process of preparing silicon carbide thin-wall tubes according to claim 1, wherein the obtaining the number of edge lines of the defect candidate region under different working conditions, calculating consistent attenuation of the development composition of the defect candidate region based on the development fluctuation working condition response dispersion and the change condition of the number of edge lines along with the working condition, specifically comprises: acquiring the number of edge lines of the same defect candidate region under at least three different forming working conditions; calculating the absolute value of the difference between the number of edge lines corresponding to each two continuous working conditions; averaging absolute values of the edge line quantity difference values of all continuous working conditions to obtain an edge line quantity change average value; Multiplying the response dispersion of the imaging fluctuation working condition of the defect candidate area by the mean value of the quantity change of the edge lines to obtain the consistent attenuation of the imaging composition of the defect candidate area.
7. The method for detecting a forming defect in a process of preparing a thin-wall tube of silicon carbide according to claim 1, wherein the steps of obtaining the main direction of texture gradient and development response intensity of the defect candidate region under different working conditions and calculating the development structure response consistency index of the defect candidate region comprise: Obtaining the main direction angle values of texture gradients of the same defect candidate region under at least three different forming working conditions; calculating the absolute value of the difference between the main direction angle values of the texture gradients corresponding to each two continuous working conditions; Averaging absolute values of the main direction angle differences of the texture gradients under all continuous working conditions to obtain a texture direction change average value; acquiring development response intensity values of the same defect candidate region under at least three different forming working conditions; Calculating the absolute value of the difference between the display response intensity values corresponding to each two continuous working conditions; Averaging absolute values of the development response intensity differences of all the continuous working conditions to obtain a development response variation average value; And multiplying the reciprocal of the mean value of the change of the texture direction by the reciprocal of the mean value of the change of the display response to obtain the display structure response consistency index of the defect candidate region.
8. The method for detecting a forming defect in a process of preparing a silicon carbide thin-wall tube according to claim 1, wherein generating a constraint factor based on a developing composition consistency attenuation index and a developing structure response consistency index, modulating a preliminary abnormal amount of the defect candidate region by using the constraint factor, and outputting a real defect detection value, specifically comprises: Calculating a constraint factor based on the display composition consistent attenuation and the display structure response consistency index, wherein the value of the constraint factor is positively correlated with the display structure response consistency index and is negatively correlated with the display composition consistent attenuation; multiplying the preliminary abnormal quantity of the defect candidate region by a constraint factor to obtain a modulated abnormal quantity; and outputting the modulated abnormal quantity as a real defect detection value of the defect candidate area.
9. The method for detecting a molding defect for a silicon carbide thin-wall tube manufacturing process according to claim 1, further comprising: based on the space position information of each defect candidate region, the defect candidate regions which are adjacent in space and the real defect detection values of which meet the preset conditions are aggregated into the same forming defect region; According to the axial position range and the circumferential angle range covered by the polymerized forming defect area, determining the space position, coverage area and contour boundary of the forming defect on the surface of the thin-wall tube; evaluating the development response intensity and the spatial stability of the formed defect based on the distribution condition of the real defect detection values of each defect candidate region in the polymerized formed defect region; And grading the severity degree or identifying the risk of the formed defect according to the evaluation result, and outputting a structural defect report, wherein the structural defect report comprises at least one of spatial position, contour boundary and grade information.
10. A molding defect detection system for use in a process for manufacturing thin wall silicon carbide tubes, comprising a memory, a processor, and a computer program stored on the memory and executable on the processor, wherein the computer program when executed by the processor performs the steps of a method for molding defect detection for use in a process for manufacturing thin wall silicon carbide tubes as claimed in any one of claims 1 to 9.

Description

Shaping defect detection method for silicon carbide thin-wall tube preparation process Technical Field The invention relates to the field of image processing, in particular to a molding defect detection method for a silicon carbide thin-wall tube preparation process. Background Silicon carbide is a typical third-generation wide-bandgap ceramic material, has the characteristics of high temperature resistance, corrosion resistance, high thermal conductivity, high mechanical strength, excellent chemical stability and the like, and therefore has wide application in the fields of high-temperature heat exchange tubes, semiconductor process pipelines, nuclear energy equipment, aerospace propulsion systems, high-end chemical reaction devices and the like. With the continuous development of high-end equipment towards light weight, miniaturization and high reliability, the silicon carbide thin-wall tube can obviously reduce the material consumption and improve the heat transfer efficiency while ensuring the structural strength, so that the silicon carbide thin-wall tube gradually becomes a key functional component in the field. However, in the process of detecting forming defects of silicon carbide thin-wall tubes, the conventional visual detection method generally faces the problem of unstable defect characterization. The method is characterized in that when a real forming defect exists in the thin-wall tube, due to the common influence of factors such as uneven microscopic density distribution, forming flow orientation difference, local thickness change and the like on the surface of the tube, the defect generated in the forming stage is often accompanied with the change of the internal stress gradient and the surface microstructure of the material, so that the defect does not show fixed gray scale, texture or edge characteristics under the condition of visible light imaging. Most of existing detection methods are based on static feature assumption or single-scale image characterization modes, and cannot establish internal correlation between the actual physical state of the forming defect and the conditional evolution of the image features of the forming defect. Therefore, the same defect often shows image characteristics with obvious differences in different imaging positions or different molding batches, so that the problems of misjudgment, missed judgment or fluctuation of a detection result along with the process working condition in the detection process are caused, and finally, the detection result lacks reliability in the aspects of space distribution and batch consistency. Disclosure of Invention The invention provides a molding defect detection method for a silicon carbide thin-wall tube manufacturing process, which aims to solve the existing problems. The invention relates to a forming defect detection method for a silicon carbide thin-wall tube manufacturing process, which adopts the following technical scheme: one embodiment of the invention provides a molding defect detection method for a silicon carbide thin-wall tube manufacturing process, which comprises the following steps: Acquiring surface images of the silicon carbide thin-wall tube under different working conditions, and recording the working condition identification and the spatial position information corresponding to each image; Dividing the surface image into at least two local subregions based on the space position information, screening abnormal subregions in the surface image, acquiring preliminary abnormal quantity of each abnormal subregion, and clustering the abnormal subregions into at least one defect candidate region; performing edge detection on each defect candidate region, obtaining edge average gradient values of all abnormal subareas in each defect candidate region, and calculating the imaging fluctuation complexity of the defect candidate region based on the edge average gradient values; Acquiring the imaging fluctuation complexity of any defect candidate region under different working conditions, and calculating the response dispersion of the imaging fluctuation working condition of the defect candidate region based on the difference of the imaging fluctuation complexity under different working conditions; Acquiring the quantity of edge lines of the defect candidate region under different working conditions, and calculating the consistent attenuation of the development composition of the defect candidate region based on the development fluctuation working condition response dispersion and the change condition of the quantity of the edge lines along with the working condition; Acquiring the main direction of texture gradient and development response intensity of the defect candidate region under different working conditions, and calculating the development structure response consistency index of the defect candidate region; Generating a constraint factor based on the consistent attenuation of the imaging composition and the