Image quality checking method for nuclear fuel rod end plug welding seam ray digital imaging defect identification

Abstract

The invention belongs to the technical field of nuclear power, and particularly relates to an image quality checking method for nuclear rod end plug weld ray digital imaging defect identification. According to the image quality checking method for nuclear fuel rod end plug welding line ray digital imaging defect identification provided by the invention, the effective to-be-evaluated image of the single nuclear fuel rod end plug can be efficiently extracted by preprocessing the image; a to-be-evaluated image of a single nuclear fuel rod is matched with a template image, and a verification result can be effectively controlled through an external threshold value mode, so that the problem of a specific fuel rod image on a group of images can be specifically positioned.

Description

Technical Field

[0001] The invention belongs to the technical field of nuclear power, and in particular relates to an image quality verification method for identifying defects in radiographic digital imaging of a nuclear fuel rod end plug weld. Background Art

[0002] With the development of computer technology, image processing, recognition, image verification and other technologies have been increasingly used. In the nuclear fuel rod production environment, before identifying the end plug weld defects, it is necessary to verify the image taken by the X-ray machine to facilitate subsequent defect detection by manual or machine vision.

[0003] Existing image verification methods usually directly perform global feature comparison between the image to be detected and the template image. However, the direct global feature comparison method cannot distinguish the subtle local differences in the image, resulting in low accuracy and efficiency of image verification.

[0004] In actual industrial environments, there are a lot of blank areas in the radiographic images of nuclear fuel rod end plug welds. At the same time, due to the limitations of some detection devices, the collected radiographic images may show the position of the fuel rods offset, making image verification more difficult and inaccurate. Summary of the invention

[0005] In order to overcome the problems existing in the related art, an image quality calibration method for defect recognition of nuclear fuel rod end plug weld by radiographic digital imaging is provided.

[0006] According to one aspect of an embodiment of the present disclosure, there is provided an image quality verification method for nuclear fuel rod end plug weld defect identification using digital radiographic imaging, the method comprising:

[0007] Step 1, collecting original images;

[0008] Step 2: Image binarization: convert all original images that are not grayscale images into grayscale images, use a Gaussian filter to smooth each grayscale image, and select a preset threshold to convert the grayscale image into a binary image.

[0009] Step 3: contour extraction: perform contour extraction on the binarized image; return a list of all contours, and return the largest outer contour in the list of all contours;

[0010] Step 4: vertical projection: vertically project the image after contour extraction, count the pixel distribution in the horizontal direction, and obtain a vertical projection image;

[0011] Step 5: single rod segmentation, obtain vertical projection image; according to the characteristics of fuel rod image, select the minimum length L of continuous pixels in the horizontal direction; in the vertical projection image, search from left to right, retain the interval where the length of continuous pixels is greater than L, discard other pixels, obtain the projection interval of the single rod area in the horizontal direction, and form a single rod segmentation image;

[0012] Step 6, image verification, preset the template images of the upper end plug and the lower end plug, and perform template matching between the formed single-stick segmentation image and the template image to extract the effective evaluation area.

[0013] In a possible implementation, in step 1, the acquired Dicom image format is converted into a 16-bit Tiff image format.

[0014] In a possible implementation, in step 2, a Gaussian binarization method is used to convert the grayscale image into a binary image.

[0015] In a possible implementation, in step 3, before performing contour extraction on the binarized image, a morphological operation is performed on the binarized image.

[0016] In a possible implementation, during the template matching process in step 6, images with any of the following conditions are screened out: missing images, no fuel rods, and incomplete insertion of fuel rods.

[0017] In one possible implementation, in step 6, the TM_CCOEFF_NORMED method is used to match the image. For the image to be matched, the image is traversed from left to right and from top to bottom, and the pixel matching degree between the template and the overlapping sub-image is calculated. The greater the matching degree, the greater the possibility of the sameness. If the image to be matched is W×H and the template image is w×h, the size of the result matrix is ​​(W-w+1)×(H-h+1); the final correlation value is scaled to the range of [-1,1], and the matching result takes the maximum value according to the returned matching matrix. If it is greater than the matching threshold, it is determined that the image matches the template.

[0018] In a possible implementation, the matching threshold is set as an adjustable parameter.

[0019] The beneficial effects of the present disclosure are as follows: the image quality verification method for nuclear fuel rod end plug weld radiographic digital imaging defect identification provided by the present disclosure can efficiently extract the effective image to be evaluated of a single nuclear fuel rod end plug by preprocessing the image; the image to be evaluated of a single nuclear fuel rod is used to match the template image, and the verification result can be effectively controlled by an external threshold method, thereby specifically locating the image problem of a specific fuel rod in a group of images. BRIEF DESCRIPTION OF THE DRAWINGS

[0020] Figure 1 The present invention is a flow chart showing an image quality verification method for defect identification of nuclear fuel rod end plug weld by radiographic digital imaging according to an exemplary embodiment.

[0021] Figure 2 To convert the image into 16-bit Tiff format;

[0022] Figure 3 The image after binarization

[0023] Figure 4 Image after contour extraction

[0024] Figure 5 The vertically projected image

[0025] Figure 6 Image after segmentation of the upper plug single rod

[0026] Figure 7 The upper plug image verification result DETAILED DESCRIPTION

[0027] The present invention is further described in detail below with reference to the accompanying drawings and specific embodiments.

[0028] Figure 1 FIG. 1 is a flow chart of an image quality verification method for nuclear fuel rod end plug weld defect identification using digital radiographic imaging according to an exemplary embodiment. Figure 1 As shown, the method includes:

[0029] Step 1: Collect the original image, such as Figure 2 As shown, the acquired Dicom image format is converted into a 16-bit Tiff image format.

[0030] Step 2, image binarization, convert all the original images that are not grayscale images into grayscale images, and use Gaussian filters to smooth each grayscale image. This helps reduce noise in the image and makes the binarization more accurate. Select a preset threshold to convert the grayscale image to a binary image. Usually, you can use adaptive thresholding to select a suitable preset threshold. Binarize the pixels of the image according to the threshold, set pixels greater than the threshold to white (255 or 1), and pixels less than the threshold to black (0). The effect is as follows Figure 3 shown.

[0031] Gaussian binarization is a local binarization method, also called adaptive threshold method, which determines the situation of a pixel by setting a region size and comparing the size relationship between this point and the average value of the pixels in the region (or other features). After using Gaussian binarization, the outline of the fuel rod image is clearer and the interference is small.

[0032] Step 3: contour extraction: perform contour extraction on the binarized image and retain the valid area in the image.

[0033] In step 3, the original image has a pixel size of 3008×2512 and has a large number of blank areas. Contour extraction is performed on the Gaussian binarized image to obtain the real compact area of ​​the image.

[0034] In a possible implementation, a morphological operation, such as dilation or erosion, may be performed on the input image as needed to better separate the target area.

[0035] Find all connected components of non-zero pixels in a binary image. Find the boundary of each connected component, i.e., the contour. Different algorithms can be used to calculate contours, such as the Suzuki algorithm or the Ramirez algorithm.

[0036] like Figure 4 As shown, a list of all contours is returned, and the largest outer contour of all contour lists is returned.

[0037] Step 4, vertical projection, vertically project the image after contour extraction, count the pixel distribution in the horizontal direction, and obtain the vertical projection image. In a possible implementation, in order to reduce noise interference, the statistical result is thresholded, that is, the number of pixels in the vertical direction less than the preset threshold is set to 0, indicating that the position is in the area separated by the rods. The effect is as follows Figure 5 shown.

[0038] Step 5: single rod segmentation to obtain a vertical projection image; according to the characteristics of the fuel rod image, the minimum length L of continuous pixels in the horizontal direction is selected.

[0039] In the vertical projection image, search from left to right, keep the intervals where pixels appear continuously for a length greater than L, discard other pixels, and obtain the horizontal projection interval of the single-bar area to form a single-bar segmentation image. Figure 6 shown.

[0040] Step 6, image verification, preset the template images of the upper and lower end plugs, and perform template matching between the formed single rod segmentation image and the template image to extract the effective assessment area. Among them, it can be checked whether the image meets the standards, whether there are any missing parts, whether the fuel rods are not fully inserted, etc.; accurately match the area to be assessed and improve the defect detection efficiency.

[0041] In step 6, the TM_CCOEFF_NORMED method (template matching method using normalized cross-correlation) can be used to traverse the image to be matched from left to right and from top to bottom, and calculate the pixel matching degree between the template and the overlapping sub-image. The greater the matching degree, the greater the possibility of being the same. If the image to be matched is W×H and the template image is w×h, the size of the result matrix is ​​(W-w+1)×(H-h+1).

[0042] The final correlation value is scaled to the range of [-1,1], and the matching result takes the maximum value according to the returned matching matrix. If it is greater than the matching threshold, the image is considered to match the template. The matching threshold is set as an adjustable parameter to increase adaptability. The final verification effect and verification similarity value are as follows: Figure 7 shown.

[0043] The embodiments of the present disclosure have been described above, and the above description is exemplary, not exhaustive, and is not limited to the disclosed embodiments. Many modifications and changes will be apparent to those of ordinary skill in the art without departing from the scope and spirit of the described embodiments. The selection of terms used herein is intended to best explain the principles of the embodiments, practical applications, or improvements to the technology in the market, or to enable other persons of ordinary skill in the art to understand the embodiments disclosed herein.

Claims

1. An image quality verification method for nuclear fuel rod end plug weld defect identification using digital radiographic imaging, characterized in that: The method comprises: Step 1, collecting original images; Step 2: Image binarization: convert all original images that are not grayscale images into grayscale images, use a Gaussian filter to smooth each grayscale image, and select a preset threshold to convert the grayscale image into a binary image. Step 3: contour extraction: perform contour extraction on the binarized image; return a list of all contours, and return the largest outer contour in the list of all contours; Step 4: vertical projection: vertically project the image after contour extraction, count the pixel distribution in the horizontal direction, and obtain a vertical projection image; Step 5: single rod segmentation, obtain vertical projection image; according to the characteristics of fuel rod image, select the minimum length L of continuous pixels in the horizontal direction; in the vertical projection image, search from left to right, retain the interval where the length of continuous pixels is greater than L, discard other pixels, obtain the projection interval of the single rod area in the horizontal direction, and form a single rod segmentation image; Step 6, image verification, preset the template images of the upper end plug and the lower end plug, and perform template matching between the formed single-stick segmentation image and the template image to extract the effective evaluation area.

2. The method according to claim 1, characterized in that In step 1, the acquired Dicom image format is converted into a 16-bit Tiff image format.

3. The method according to claim 1, characterized in that In step 2, the grayscale image is converted into a binary image using the Gaussian binarization method.

4. The method according to claim 1, characterized in that In step 3, before performing contour extraction on the binarized image, a morphological operation is performed on the binarized image.

5. The method according to claim 1, characterized in that During the template matching process in step 6, images with any of the following conditions are screened out: missing images, no fuel rods, and incomplete insertion of fuel rods.

6. The method according to claim 1, characterized in that In step 6, the TM_CCOEFF_NORMED method is used to match the image. For the image to be matched, the image is traversed from left to right and from top to bottom, and the pixel matching degree between the template and the overlapping sub-image is calculated. The greater the matching degree, the greater the possibility of the same. If the image to be matched is W×H and the template image is w×h, the size of the result matrix is ​​(W-w+1)×(H-h+1); the final correlation value is scaled to the range of [-1,1], and the matching result takes the maximum value according to the returned matching matrix. If it is greater than the matching threshold, it is determined that the image matches the template.

7. The method according to claim 1, characterized in that The matching threshold is set as an adjustable parameter.

Images