Automatic Identification Method and Device for Defects on the Outer Surface of a Nuclear Power Plant Containment

Abstract

The present invention provides a method and device for automatically identifying defects on the outer surface of a containment vessel of a nuclear power plant. The method includes: acquiring a to-be-detected image of the outer surface of the containment vessel; processing the to-be-detected image to obtain a shadow-removed image of the outer surface of the containment vessel; enhancing defect features in the shadow-removed image and obtaining a local image with enhanced defect features; identifying defect pixel points in the local image and obtaining a defect pixel point coordinate map by plotting according to the positions of each defect pixel point; and obtaining an identification result of the defects on the outer surface of the containment vessel according to the defect pixel point coordinate map. The present invention is used to solve problems such as high industrial risks, easy missed detection and false detection in traditional manual detection of defects on the outer surface of a nuclear power plant containment vessel, and can timely detect defects on the outer surface of the containment vessel, thereby improving the safety performance of the containment vessel.

Description

Technical Field

[0001] The present invention belongs to the technical field of nuclear power, and particularly relates to a method and device for automatically identifying defects on the outer surface of a nuclear power plant containment vessel. Background Art

[0002] The containment vessel is one of the important barriers for a nuclear power plant to prevent the leakage of fission products, and it is required to provide good sealing performance under the design basis accident temperature and pressure conditions. Cracks are an early manifestation of damage to the nuclear power plant containment vessel and are also an important indicator for evaluating the quality of the wall. When a steam pipe rupture or a loss-of-coolant accident occurs in a nuclear power plant, high-temperature and high-pressure steam is released, causing the temperature and pressure inside the containment vessel to rise. If there are defects in the containment vessel, it may lead to the failure of its function of containing radioactivity under accident conditions due to its failure to meet the design requirements, thus triggering a major safety accident. Therefore, the performance of the containment vessel directly affects the accident handling ability of the nuclear power plant.

[0003] At present, the main method for collecting defect data on the outer surface of the containment vessel is manual inspection, which has problems such as high industrial risk, a relatively high probability of missed detection, and large subjective judgment errors of people. Summary of the Invention

[0004] The technical problem to be solved by the present invention is to provide a method and device for automatically identifying defects on the outer surface of a nuclear power plant containment vessel in view of the above deficiencies in the prior art, so as to solve the problems of high industrial risk, easy missed detection and misdetection in traditional manual detection, and be able to timely detect defects on the outer surface of the containment vessel, thereby improving the safety performance of the containment vessel.

[0005] To solve the above technical problems, the present invention adopts the following technical solutions:

[0006] In the first aspect of the present invention, there is provided a method for automatically identifying defects on the outer surface of a nuclear power plant containment vessel, the method comprising:

[0007] Obtaining a to-be-detected image of the outer surface of the containment vessel;

[0008] Processing the to-be-detected image to obtain a shadow-removed image of the outer surface of the containment vessel;

[0009] Enhancing defect features in the shadow-removed image and obtaining a local image with enhanced defect features;

[0010] Identifying defect pixel points in the local image and obtaining a defect pixel point coordinate map by drawing according to the positions of each defect pixel point;

[0011] Obtaining an identification result of the defects on the outer surface of the containment vessel according to the defect pixel point coordinate map.

[0012] Preferably, the obtaining of the image to be detected on the outer surface of the containment vessel specifically includes:

[0013] Dividing the outer surface of the containment vessel into multiple regions;

[0014] Performing remote sensing photography on the multiple regions respectively to obtain multiple remote sensing images;

[0015] Stitching the multiple remote sensing images according to the same size ratio to obtain the overall stitched image of the area to be detected on the outer surface of the containment vessel;

[0016] Performing grayscale processing on the overall stitched image to obtain the image to be detected on the outer surface of the containment vessel.

[0017] Preferably, the processing of the image to be detected to obtain the shadow removal image on the outer surface of the containment vessel specifically includes:

[0018] Using the minimum-maximum filtering method based on a pixel window of a selected size to filter the image to be detected to obtain the shadow removal image on the outer surface of the containment vessel;

[0019] The enhancing of the defect features in the shadow removal image and obtaining the local image with enhanced defect features specifically includes:

[0020] Using an edge detection algorithm to perform edge detection and enhancement on the shadow removal image to obtain an edge enhanced image containing the image edges representing the defect features;

[0021] Cropping the local image where each image edge in the edge enhanced image is located to obtain several local images with enhanced defect features;

[0022] The identifying of the defective pixel points in the local image and drawing the coordinate map of the defective pixel points according to the positions of each defective pixel point specifically includes:

[0023] Using a corner detection algorithm to detect the corners in each local image and taking the corners as the defective pixel points;

[0024] Establishing a coordinate map consistent with the ratio of each local image and drawing the defective pixel points in the coordinate map according to their positions in the corresponding local image to obtain the coordinate map of the defective pixel points.

[0025] Preferably, before using the minimum-maximum filtering method based on a pixel window of a selected size to filter the image to be detected, the method further includes:

[0026] Based on the basic pixels of the image to be detected, presetting several pixel windows of different sizes;

[0027] When performing the first automatic defect identification on the outer surface of the containment, the minimum-maximum filtering method based on pixel windows of different preset sizes is respectively selected to filter the to-be-detected image obtained for the first time;

[0028] The pixel window with the maximum preset size that meets the processing requirements of the preset shadow removal effect after filtering is determined as the pixel window with the selected size.

[0029] Preferably, the edge detection algorithm is used to perform edge detection and enhancement on the shadow removal image to obtain an edge-enhanced image containing the image edges representing defect features, specifically:

[0030] The following formula is used to process the shadow removal image to obtain the edge-enhanced image:

[0031]

[0032] where: P is the edge-enhanced image, P0 is the shadow removal image, is the x-axis operator, is the y-axis operator.

[0033] Preferably, the corner detection algorithm is specifically the Harris algorithm.

[0034] Preferably, obtaining the identification result of the defect on the outer surface of the containment according to the defect pixel point coordinate map specifically includes:

[0035] Connect the defect pixel points with a distance less than the preset value in the defect pixel point coordinate map to form a defect pixel point connection diagram, and the defect pixel point connection diagram corresponds to the defect on the outer surface of the containment;

[0036] According to the defect pixel point coordinate map and the defect pixel point connection diagram, and in combination with the position and proportional relationship corresponding to the outer surface of the containment, obtain the shape, position and size of the defect on the outer surface of the containment.

[0037] Preferably, after obtaining the shape, position and size of the defect on the outer surface of the containment, the method further includes:

[0038] Enter the shape, position and size of the defect on the outer surface of the containment into the database.

[0039] In the second aspect of the present invention, an automatic defect identification device for the outer surface of a nuclear power plant containment is provided, and the device includes:

[0040] An acquisition module for acquiring the to-be-detected image of the outer surface of the containment;

[0041] A processing module, connected to the acquisition module, for processing the image to be detected to obtain a shadow-removed image of the outer surface of the containment;

[0042] An enhancement module, connected to the processing module, for enhancing the defect features in the shadow-removed image and obtaining a local image with enhanced defect features;

[0043] A first recognition module, connected to the enhancement module, for identifying the defective pixel points in the local image and obtaining a coordinate map of the defective pixel points by drawing according to the position of each defective pixel point;

[0044] A second recognition module, connected to the first recognition module, for obtaining the recognition result of the defect on the outer surface of the containment according to the coordinate map of the defective pixel points.

[0045] Preferably, the acquisition module includes:

[0046] A division unit for dividing the outer surface of the containment into multiple regions;

[0047] A remote sensing photographing unit, connected to the division unit, for respectively performing remote sensing photographing on multiple regions to obtain multiple remote sensing images;

[0048] A splicing unit, connected to the remote sensing photographing unit, for splicing multiple remote sensing images according to the same size ratio to obtain an overall spliced image of the area to be detected on the outer surface of the containment;

[0049] A grayscale processing unit, connected to the splicing unit, for performing grayscale processing on the overall spliced image to obtain the image to be detected on the outer surface of the containment.

[0050] The automatic defect recognition method and device for the outer surface of the nuclear power plant containment provided by the present invention realizes the automatic recognition of the defects on the outer surface of the containment based on the digital image processing method. By removing the shadow, locally enhancing the defect features, and then detecting the local defect features to determine the defects, the defect recognition result of the outer surface of the containment is obtained. The defect recognition effect of the present invention is good, and it can be applied to large-sized containments of different reactor types including pressurized water reactor nuclear power plants, and can timely detect the defects on the outer surface of the containment, thereby improving the safety performance of the containment. Description of the Drawings

[0051] Figure 1 It is a flowchart of the automatic defect recognition method for the outer surface of the nuclear power plant containment in Embodiment 1 of the present invention;

[0052] Figure 2 It is a flowchart of the automatic defect recognition method for the outer surface of the nuclear power plant containment in another embodiment of the present invention;

[0053] Figure 3 Flow chart of the method for selecting the pixel window size in the minimum-maximum filtering method based on a pixel window in another embodiment of the present invention;

[0054] Figure 4 Image to be detected on the outer surface of the containment vessel of a nuclear power plant in another embodiment of the present invention;

[0055] Figure 5 Shadow removal image obtained by using the minimum-maximum filtering method based on pixel windows of different preset sizes in another embodiment of the present invention; In the figure: (a), (b), (c), and (d) are shadow removal images obtained by the minimum-maximum filtering method based on pixel windows of 10, 4, 2, and 1 times the basic pixel unit, respectively;

[0056] Figure 6 Shadow removal image obtained by using the minimum-maximum filtering method based on the selected pixel window size in another embodiment of the present invention;

[0057] Figure 7 Local image with enhanced defect features in another embodiment of the present invention;

[0058] Figure 8 Coordinate map of defective pixel points in another embodiment of the present invention;

[0059] Figure 9 Schematic structural diagram of the automatic defect recognition device for the outer surface of the containment vessel of a nuclear power plant in Embodiment 2 of the present invention. Detailed implementation manners

[0060] Next, the technical solutions in the present invention will be clearly and completely described in conjunction with the accompanying drawings in the present invention. Obviously, the described embodiments are part of the embodiments of the present invention, rather than all of the embodiments. All other embodiments obtained by those of ordinary skill in the art based on the embodiments in the present invention without creative efforts shall fall within the scope of the present invention.

[0061] In the description of the present invention, it should be noted that the orientation or positional relationship indicated by terms such as "upper" is based on the orientation or positional relationship shown in the accompanying drawings. It is only for the convenience and simplification of description, rather than indicating or implying that the device or element referred to must have a specific orientation, be constructed and operated in a specific orientation, and therefore should not be construed as a limitation to the present invention.

[0062] In the description of the present invention, the terms "first" and "second" are only used for descriptive purposes and cannot be construed as indicating or implying relative importance.

[0063] In the description of the present invention, it should be noted that unless otherwise clearly specified and defined, terms such as "connection", "installation", "fixation", etc. should be understood in a broad sense. For example, it can be a fixed connection, a detachable connection, or an integral connection; it can be directly connected, or indirectly connected through an intermediate medium, and it can be the communication inside two components. For those skilled in the art, the specific meanings of the above terms in the present invention can be understood according to specific situations.

[0064] In the description of the present invention, each unit and module involved may correspond to only one entity structure, or may be composed of multiple entity structures. Alternatively, multiple units and modules may also be integrated into one entity structure; the units and modules involved can be implemented in software or in hardware. For example, the units and modules can be located in the processor.

[0065] In the description of the present invention, without conflict, the functions and steps marked in the flowcharts and block diagrams of the present invention may occur in an order different from that marked in the drawings.

[0066] Embodiment 1:

[0067] As Figure 1 shown, Embodiment 1 of the present invention provides an automatic defect recognition method for the outer surface of a nuclear power plant containment. The method includes:

[0068] S11. Obtain a to-be-detected image of the outer surface of the containment.

[0069] In this embodiment, an automatic defect recognition method for the outer surface of a nuclear power plant containment based on digital image processing is given to solve problems such as high industrial risk, easy omission and misdetection in traditional manual detection. Therefore, in order to identify the defects on the outer surface of the containment, it is first necessary to obtain the outer surface image of the containment. The method for obtaining the outer surface image of the containment can use various existing imaging technical means to obtain the outer surface image features of the containment, and no specific limitation is made here.

[0070] In an optional embodiment, the obtaining of the to-be-detected image of the outer surface of the containment specifically includes:

[0071] Divide the outer surface of the containment into multiple regions;

[0072] Conduct remote sensing shooting on the multiple regions respectively to obtain multiple remote sensing images;

[0073] Stitch the multiple remote sensing images according to the same size ratio to obtain an overall stitched image of the to-be-detected area on the outer surface of the containment;

[0074] Perform grayscale processing on the overall stitched image to obtain the to-be-detected image of the outer surface of the containment.

[0075] In this embodiment, as Figure 2 shown, the specific steps for obtaining the image to be detected on the outer surface of the containment vessel include: S001 remote sensing photographing of the outer surface of the containment vessel, and S002 image stitching and grayscale processing; specifically, a remote sensing camera and a computer can be used to implement steps S001 and S002 respectively. First, an image of the outer surface of the containment vessel is obtained by using the method of remote sensing photographing. Since the containment vessel of a nuclear power plant is huge, with an inner diameter usually exceeding 20 meters, the outer surface image of the containment vessel is usually obtained by stitching images taken at different positions. Specifically, different outer surface images are obtained by remote slicing, and after correcting the size ratios of different outer surface images, the overall image is stitched. After obtaining the overall image of the area to be detected on the outer surface of the containment vessel, the original image is preprocessed according to the recognition purpose to exclude unnecessary image information, thereby reducing the computational amount of subsequent recognition. Mainly, it is the grayscale processing of converting the three-channel signal of the original image into a single-channel signal, etc.

[0076] In a more specific embodiment, the automatic recognition of defects on the outer surface of the containment vessel of a certain pressurized water reactor nuclear power plant is described. The elevation inside the containment vessel building of this pressurized water reactor nuclear power plant is about 40m, and the outer surface area is about 7000m 2 , by pre-dividing the outer surface of the containment vessel into multiple regions, and then performing sub-region remote sensing photographing, overall image stitching and grayscale processing, the Figure 4 image to be detected as shown is obtained. Figure 4 The image to be detected in

[0077] S12. Process the image to be detected to obtain a shadow-removed image of the outer surface of the containment vessel.

[0078] In this embodiment, the image to be detected needs to be denoised. Especially, the outer surface images of the containment vessel are all obtained outdoors. There will inevitably be light effects and interference factors such as shadows at the outdoor surveying and mapping site, which will cause large errors in the extraction of the crack edge contour. If not eliminated, it will inevitably affect the accuracy and reliability of defect recognition. Therefore, it needs to be processed to eliminate the shadows.

[0079] In an alternative embodiment, the processing of the image to be detected to obtain a shadow-removed image of the outer surface of the containment vessel specifically includes:

[0080] Filter the image to be detected by using the minimum-maximum filtering method based on a pixel window of a selected size to obtain a shadow-removed image of the outer surface of the containment vessel.

[0081] In this embodiment, asFigure 2 As shown, for the outdoor shooting of the image of the outer surface of the containment of a nuclear power plant, considering the characteristics that its shadow has a larger gray value than the expected information to be retained, the method of S003 image filtering is adopted to remove the shadow on the outer surface of the containment. Considering that inappropriate shadow processing methods will overly reduce the defect features and also decrease the accuracy of defect recognition, it is necessary to select a suitable filtering algorithm. At the same time, according to existing experience, selecting a suitable filtering window for the same filtering algorithm can obtain a better filtering effect. In the prior art, the relevant filtering functions include: high-pass filtering, low-pass filtering, Gaussian difference band-pass filtering, mean filtering, maximum-minimum filtering, etc. Among them, maximum-minimum filtering is a filtering method that traverses the entire digital image and replaces each pixel with the maximum and minimum gray values in the surrounding area. Maximum-minimum filtering is a relatively conservative filtering method. All of the above filtering methods can be used in the present invention. After comparative experiments, the pixel window size with good shadow removal effect and complete retention of the image defect feature information is preselected in advance, and the shadow removal effect of the image of the outer surface of the containment completed based on the maximum-minimum filtering function is the best. Therefore, the maximum-minimum filtering method based on the pixel window of the selected size is used to filter the to-be-detected image.

[0082] In an alternative embodiment, before using the maximum-minimum filtering method based on the pixel window of the selected size to filter the to-be-detected image, the method further includes:

[0083] Based on the basic pixels of the to-be-detected image, a plurality of pixel windows of different sizes are preset;

[0084] When performing the first automatic defect recognition on the outer surface of the containment, the maximum-minimum filtering methods based on the pixel windows of different preset sizes are respectively selected to filter the to-be-detected image obtained for the first time;

[0085] The pixel window with the largest preset size that meets the preset shadow removal effect processing requirements after filtering is determined as the pixel window of the selected size.

[0086] In this embodiment, since there are differences in the shadow removal effects of pixel windows of different sizes, the shadow removal effect of a large window is inferior to that of a small window. Although the processing of an overly small pixel window is fine and the shadow removal effect is good, it also causes the loss of the main feature information of defects such as cracks. Therefore, as Figure 3As shown, after determining the photographing points for each area of the containment vessel outer surface image, a pixel window of a selected size is obtained through the following steps: S121 Obtain the image to be detected on the containment vessel outer surface for the first time; S122 Select a pixel window of a preset size in descending order; S123 Perform minimum filtering on the image to be detected; S124 Perform maximum filtering on the image to be detected; S125 Analyze whether the filtering result meets the processing requirements. If not, return to step S122, re-select another pixel window of a preset size, and repeat steps S123 - S125. If so, perform step S126 to obtain a pixel window of the selected size. Among them, the processing requirements in step S125 can be determined by judging the shadow removal effect of known features such as pipelines on the outer surface images collected under different lighting conditions. That is, under certain lighting conditions, the position and size of the shadows projected by known pipelines on the containment vessel outer surface are also known. By judging whether the shadows are removed after image filtering, the shadow removal effect can be determined, and at the same time, the window with the largest size that can remove the shadows is determined as the optimal pixel window to reduce the loss of defect feature information. In subsequent regular monitoring, image filtering processing will be completed based on this pixel window to remove the image shadows.

[0087] In a more specific embodiment, 10, 4, 2, and 1 times the basic pixel unit window are respectively preset, and minimum-maximum filtering is respectively performed based on the pixel windows of the above four sizes Figure 4 on the image to be detected, and the processing results are as Figure 5 shown. Among them, the images filtered by the pixel windows of 10, 4, and 2 times the size do not meet the shadow removal effect. Therefore, the pixel window of 1 times the size is selected as the reference window for shadow removal of the image to be detected on the containment vessel outer surface of this project, and finally the shadow removal image as shown in Figure 6 is obtained.

[0088] S13. Enhance the defect features in the shadow removal image and obtain a local image with enhanced defect features.

[0089] In this embodiment, to avoid the decline of defect features in the image after removing the shadow interference factors, a suitable method needs to be adopted to enhance the defect data to improve the accuracy of defect recognition, and a local image where the defect features are located is obtained from the image with enhanced defect features to improve the efficiency of defect recognition.

[0090] In an alternative embodiment, the enhancing the defect features in the shadow removal image and obtaining a local image with enhanced defect features specifically includes:

[0091] Performing edge detection and enhancement on the shadow removal image using an edge detection algorithm to obtain an edge-enhanced image containing the image edges representing the defect features;

[0092] Intercept the local image where each image edge in the edge-enhanced image is located to obtain a number of the locally enhanced images of the defect features.

[0093] In this embodiment, since the edge of an image is an important feature of the image, especially for the defects on the outer surface of the containment vessel, one of the direct characterizations of a crack is the image edge. Therefore, as Figure 2 shown, the defect feature enhancement of the outer surface of the containment vessel is achieved through S004 edge detection.

[0094] In an alternative embodiment, the edge detection algorithm is used to perform edge detection and enhancement on the shadow-removed image to obtain an edge-enhanced image containing the image edges representing the defect features. Specifically:

[0095] The following formula is used to process the shadow-removed image to obtain the edge-enhanced image:

[0096]

[0097] where: P is the edge-enhanced image, P0 is the shadow-removed image, is the x-axis operator, is the y-axis operator.

[0098] In this embodiment, since the edges of actual images are often a combination of various types of edges and their blurred results, and there is noise in the actual image signal, both this noise and the image edges belong to high-frequency signals, and it is difficult to make a trade-off through frequency bands. Therefore, how to detect edges is a difficult problem in image processing. Commonly used edge detection operators include first-order operators, second-order Laplacian operators, and zero-crossing operators, etc. Although in principle, edge detection algorithms can apply higher-order derivatives, due to the influence of noise, in actual applications, usually only the first-order or second-order derivatives are used, and sensitivity to noise will occur when using a single second-order derivative. Through the analysis of practical results, it is determined that the edge detection algorithm based on the first-order operator can achieve defect feature enhancement in the spliced image of the outer surface of the containment vessel after shadow removal. By convolving the image with operators in the x and y axis directions, the image after edge detection enhancement is obtained. The practical results prove that the first-order operator image enhancement method can achieve defect feature enhancement for the large-size outer surface image of the nuclear power plant containment vessel.

[0099] In a more specific embodiment, after performing first-order operator image enhancement on the Figure 6 shadow-removed image described above, a locally enhanced image of a crack defect feature on the outer surface of the containment vessel obtained by interception is as Figure 7 shown.

[0100] S14. Identify the defective pixel points in the local image, and draw a defective pixel point coordinate map based on the positions of each defective pixel point.

[0101] In this embodiment, for concrete surfaces such as the outer surface of the containment, the difference between the crack-free image and the cracked defective image can essentially be regarded as different image features. The identification of defects can be achieved by identifying the defective pixel points in the image, and after identifying the defects, the clear positioning of the defects can be further achieved by plotting coordinates.

[0102] In an alternative embodiment, the identifying the defective pixel points in the local image and drawing a defective pixel point coordinate map based on the positions of each defective pixel point specifically includes:

[0103] Use a corner detection algorithm to detect the corners in each local image, and take the corners as the defective pixel points;

[0104] Establish a coordinate map consistent with the scale of each local image, and plot the defective pixel points in the coordinate map according to their positions in the corresponding local image to obtain the defective pixel point coordinate map.

[0105] In this embodiment, as Figure 2 shown, the method of introducing S005 corner detection is used to detect the defects on the outer surface of the containment. The principle of this method is to analyze the displacements of a certain area of the image in four directions: vertical, horizontal, positive diagonal, and negative diagonal, and calculate the sum of the squared differences of the correlations of each pixel point. The larger this value, the worse the correlation, and the more likely the central pixel point is to be a corner. Since the defective information has a poor correlation with the normal surface information around the defect, the purpose of defect identification can be achieved by identifying corners. Corner detection can meet the requirements for detecting defects on the large-sized outer surface of the containment in nuclear power plants, and is applicable to different reactor types with large-sized containments designed, including pressurized water reactor nuclear power plants.

[0106] In an alternative embodiment, the corner detection algorithm is specifically the Harris algorithm.

[0107] In this embodiment, it is necessary to determine a suitable digital image recognition method for the defects on the outer surface of the containment. The practical results prove that the Harris algorithm can obtain satisfactory recognition results. The Harris algorithm is a prior art and will not be elaborated here.

[0108] In a more specific embodiment, the local image shown in Figure 7 is subjected to corner detection using the Harris algorithm. The detection results are shown in Figure 8, and several corners that may be defective pixel points are detected and plotted in the coordinate map correspondingly for subsequent defect positioning.

[0109] S15. Obtain the recognition result of the defect on the outer surface of the containment vessel according to the defect pixel coordinate map.

[0110] In this embodiment, after obtaining the defect pixel coordinate map, as Figure 2 shown, finally, S006 is also required to obtain the defect recognition result. Therefore, defect localization is achieved through the coordinate map, and the localization result is used as the defect recognition result of the outer surface of the containment vessel this time.

[0111] In an alternative embodiment, the obtaining the recognition result of the defect on the outer surface of the containment vessel according to the defect pixel coordinate map specifically includes:

[0112] Connect the defect pixels with a distance less than a preset value in the defect pixel coordinate map to form a defect pixel connection diagram, and the defect pixel connection diagram corresponds to the defect on the outer surface of the containment vessel;

[0113] According to the defect pixel coordinate map and the defect pixel connection diagram, and in combination with its corresponding position and proportional relationship on the outer surface of the containment vessel, obtain the shape, position, and size of the defect on the outer surface of the containment vessel.

[0114] In this embodiment, for a defect form such as a crack, the defect pixels with a distance less than a certain value from each other can be connected to restore the crack characteristics, and then the shape, position, size, and other characteristics of the crack on the outer surface of the containment vessel can be determined according to the coordinates to obtain the defect recognition result.

[0115] In a more specific embodiment, according to Figure 8 the corner defect detection result in, the shape of the crack is shown according to the shape of the connection of the defect pixels in the figure, and the specific position and size of the defect are determined by combining the position of the image corresponding to the containment vessel and the image size ratio, that is, the defect is recognized.

[0116] In an alternative embodiment, after obtaining the shape, position, and size of the defect on the outer surface of the containment vessel, the method further includes:

[0117] Enter the shape, position, and size of the defect on the outer surface of the containment vessel into the database.

[0118] In this embodiment, by entering the shape, position, and size of the defect on the outer surface of the containment vessel into the database, the detection of the defect on the outer surface of the containment vessel for this image is completed, providing a data basis for counting the defects identified on the outer surface of the containment vessel, and then distinguishing the degree of danger of the defects to take measures in advance to control the harm caused by the defects and ensure the safety performance of the containment vessel.

[0119] Embodiment 2:

[0120] AsFigure 9 As shown in Figure 9 , Embodiment 2 of the present invention provides an automatic recognition device for defects on the outer surface of a nuclear power plant containment vessel. The device includes:

[0121] An acquisition module 11, configured to acquire a to-be-detected image of the outer surface of the containment vessel;

[0122] A processing module 12, connected to the acquisition module 11, configured to process the to-be-detected image to obtain a shadow-removed image of the outer surface of the containment vessel;

[0123] An enhancement module 13, connected to the processing module 12, configured to enhance defect features in the shadow-removed image and obtain a local image with enhanced defect features;

[0124] A first recognition module 14, connected to the enhancement module 13, configured to identify defect pixel points in the local image and draw a coordinate map of defect pixel points based on the positions of each defect pixel point;

[0125] A second recognition module 15, connected to the first recognition module 14, configured to obtain an identification result of the defect on the outer surface of the containment vessel according to the coordinate map of defect pixel points.

[0126] In an alternative embodiment, the acquisition module 11 specifically includes:

[0127] A division unit, configured to divide the outer surface of the containment vessel into multiple regions;

[0128] A remote sensing shooting unit, connected to the division unit, configured to perform remote sensing shooting on multiple regions respectively to obtain multiple remote sensing images;

[0129] A splicing unit, connected to the remote sensing shooting unit, configured to splice multiple remote sensing images according to the same size ratio to obtain an overall spliced image of the to-be-detected area on the outer surface of the containment vessel;

[0130] A grayscale processing unit, connected to the splicing unit, configured to perform grayscale processing on the overall spliced image to obtain the to-be-detected image of the outer surface of the containment vessel.

[0131] In an alternative embodiment, the processing module 12 specifically includes:

[0132] A filtering unit, configured to filter the to-be-detected image by using a minimum-maximum filtering method based on a pixel window of a selected size to obtain the shadow-removed image of the outer surface of the containment vessel;

[0133] The enhancement module 13 specifically includes:

[0134] An edge detection unit, which is used to perform edge detection and enhancement on the shadow-removed image by using an edge detection algorithm, so as to obtain an edge-enhanced image containing image edges representing defect features;

[0135] A local cropping unit, connected to the edge detection unit, which is used to crop the local image where each image edge in the edge-enhanced image is located, so as to obtain a plurality of local images with enhanced defect features;

[0136] The first recognition module 14 specifically includes:

[0137] A corner detection unit, which is used to detect corners in each local image by using a corner detection algorithm and take the corners as the defective pixel points;

[0138] A coordinate graph drawing unit, connected to the corner detection unit, which is used to establish a coordinate graph consistent with the scale of each local image and draw the defective pixel points in the coordinate graph according to their positions in the corresponding local image, so as to obtain the defective pixel point coordinate graph.

[0139] In an optional embodiment, the processing module 12 further includes:

[0140] A preset unit, which is used to preset a plurality of pixel windows with different sizes based on the basic pixels of the image to be detected;

[0141] A first filtering unit, connected to the preset unit, which is used to perform filtering on the initially obtained image to be detected by respectively selecting the minimum-maximum filtering method based on pixel windows with different preset sizes when performing automatic recognition of initial defects on the outer surface of the containment;

[0142] A selection unit, connected to the first filtering unit, which is used to determine the pixel window with the largest preset size that meets the preset shadow removal effect processing requirements after filtering as the selected size pixel window.

[0143] In an optional embodiment, the edge detection unit specifically:

[0144] Process the shadow-removed image by using the following formula to obtain the edge-enhanced image:

[0145]

[0146] where: P is the edge-enhanced image, P0 is the shadow-removed image, is the x-axis operator, is the y-axis operator.

[0147] In an optional embodiment, the corner detection unit specifically uses the Harris algorithm to detect corners in each local image.

[0148] In an alternative embodiment, the second recognition module 15 specifically includes:

[0149] A connection unit, configured to connect defective pixel points with a distance less than a preset value in the defective pixel point coordinate map to form a defective pixel point connection map, where the defective pixel point connection map corresponds to the defect on the outer surface of the containment;

[0150] A positioning unit, connected to the connection unit, configured to obtain the shape, position, and size of the defect on the outer surface of the containment according to the defective pixel point coordinate map and the defective pixel point connection map, and in combination with the position and proportional relationship corresponding to the outer surface of the containment.

[0151] In an alternative embodiment, the second recognition module 15 further includes:

[0152] An input unit, connected to the positioning unit, configured to input the shape, position, and size of the defect on the outer surface of the containment into a database.

[0153] The method and device for automatically recognizing defects on the outer surface of a containment of a nuclear power plant provided in Embodiments 1-2 of the present invention realize the automatic recognition of defects on the outer surface of the containment based on a digital image processing method. By removing shadows from the image, locally enhancing defect features, and then detecting the local defect features to determine the defects, the defect recognition result of the outer surface of the containment is obtained, so that the defect recognition effect of the present invention is good, and it can be applied to large-sized containments of different reactor types including pressurized water reactor nuclear power plants, and can timely detect defects on the outer surface of the containment, thereby improving the safety performance of the containment.

[0154] It can be understood that the above embodiments are merely exemplary embodiments adopted to illustrate the principles of the present invention, and the present invention is not limited thereto. For those of ordinary skill in the art, various modifications and improvements can be made without departing from the spirit and essence of the present invention, and these modifications and improvements are also considered within the protection scope of the present invention.

Claims

1. An automatic recognition method for defects on the outer surface of a nuclear power plant containment vessel, characterized in that The defect is a crack, and the method includes: Obtaining a to-be-detected image of the outer surface of the containment vessel; Selecting a filtering algorithm to process the to-be-detected image to obtain a shadow-removed image of the outer surface of the containment vessel. The filtering algorithm filters the to-be-detected image based on a pixel window of a selected size, and the pixel window of the selected size is the window with the largest size for which the known shadow of the known feature in the to-be-detected image is filtered out; Enhancing the defect features in the shadow-removed image and obtaining a local image with enhanced defect features, specifically including: Performing edge detection and enhancement on the shadow-removed image by using an edge detection algorithm to obtain an edge-enhanced image containing the image edges representing the defect features; Cropping the local image where each image edge in the edge-enhanced image is located to obtain a plurality of local images with enhanced defect features; Identifying the defective pixel points in the local image and drawing a defective pixel point coordinate map according to the positions of each defective pixel point; Connecting the lines according to the defective pixel point coordinate map to obtain the recognition result of the defect on the outer surface of the containment vessel.

2. The method according to claim 1, characterized in that, The obtaining of the to-be-detected image of the outer surface of the containment vessel specifically includes: Dividing the outer surface of the containment vessel into multiple regions; Performing remote sensing shooting on the multiple regions respectively to obtain a plurality of remote sensing images; Stitching the plurality of remote sensing images according to the same size ratio to obtain an overall stitched image of the to-be-detected area on the outer surface of the containment vessel; Performing grayscale processing on the overall stitched image to obtain the to-be-detected image of the outer surface of the containment vessel.

3. The method according to claim 1 or 2, characterized in that, The processing of the to-be-detected image to obtain the shadow-removed image of the outer surface of the containment vessel specifically includes: Filtering the to-be-detected image by using the minimum-maximum filtering method based on a pixel window of a selected size to obtain the shadow-removed image of the outer surface of the containment vessel; The identifying of the defective pixel points in the local image and the drawing of the defective pixel point coordinate map according to the positions of each defective pixel point specifically includes: Detecting the corner points in each local image by using a corner point detection algorithm and taking the corner points as the defective pixel points; Establishing a coordinate map consistent with the ratio of each local image and drawing the defective pixel points in the coordinate map according to their positions in the corresponding local image to obtain the defective pixel point coordinate map.

4. The method according to claim 3, characterized in that, Before filtering the to-be-detected image by using the minimum-maximum filtering method based on a pixel window of a selected size, the method further includes: Presetting a plurality of pixel windows of different sizes based on the basic pixels of the to-be-detected image; When performing the first automatic defect recognition on the outer surface of the containment vessel, respectively selecting the minimum-maximum filtering method based on pixel windows of different preset sizes to filter the to-be-detected image obtained for the first time; Determining the pixel window with the largest preset size that meets the preset shadow-removing effect processing requirements after filtering as the pixel window of the selected size. The preset shadow-removing effect processing requirements include that the known shadow of the known features on the outer surface of the containment vessel under certain lighting conditions is filtered out.

5. The method according to claim 3, characterized in that, The edge detection algorithm is used to perform edge detection and enhancement on the shadow-removed image to obtain an edge-enhanced image containing the image edges representing the defect features. Specifically: The following formula is used to process the shadow-removed image to obtain the edge-enhanced image: Where: P is the edge-enhanced image, and P0 is the shadow-removed image. is the x-axis operator. is the y-axis operator.

6. The method according to claim 3, wherein The corner detection algorithm is specifically the Harris algorithm.

7. The method according to claim 3, characterized in that Connecting the lines according to the defect pixel point coordinate map to obtain the identification result of the defect on the outer surface of the containment vessel specifically includes: Connecting the defect pixel points with a distance less than a preset value in the defect pixel point coordinate map to form a defect pixel point connection map, and the defect pixel point connection map corresponds to the defect on the outer surface of the containment vessel; Based on the defect pixel point coordinate map and the defect pixel point connection map, and combined with its corresponding position and proportional relationship on the outer surface of the containment vessel, the shape, position, and size of the defect on the outer surface of the containment vessel are obtained.

8. The method according to claim 7, wherein After obtaining the shape, position, and size of the defect on the outer surface of the containment vessel, the method further includes: Entering the shape, position, and size of the defect on the outer surface of the containment vessel into the database.

9. An automatic recognition device for defects on the outer surface of a nuclear power plant containment vessel, characterized in that, The defect is a crack, and the device includes: An acquisition module for acquiring the image to be detected on the outer surface of the containment vessel; A processing module connected to the acquisition module for selecting a filtering algorithm to process the image to be detected to obtain the shadow-removed image of the outer surface of the containment vessel. The filtering algorithm filters the image to be detected based on a pixel window of a selected size, and the pixel window of the selected size is the largest size window for filtering the known feature shadows in the image to be detected; An enhancement module connected to the processing module for enhancing the defect features in the shadow-removed image and obtaining a local image with enhanced defect features. Specifically includes: An edge detection unit for performing edge detection and enhancement on the shadow-removed image using an edge detection algorithm to obtain an edge-enhanced image containing the image edges representing the defect features, A local intercepting unit connected to the edge detection unit for intercepting the local image where each image edge in the edge-enhanced image is located to obtain a plurality of local images with enhanced defect features; A first identification module connected to the enhancement module for identifying the defect pixel points in the local image and obtaining a defect pixel point coordinate map by drawing according to the position of each defect pixel point; A second identification module connected to the first identification module for connecting the lines according to the defect pixel point coordinate map to obtain the identification result of the defect on the outer surface of the containment vessel.

10. The device according to claim 9, wherein, The acquisition module includes: A dividing unit for dividing the outer surface of the containment vessel into multiple regions; A remote sensing shooting unit connected to the dividing unit for performing remote sensing shooting on the multiple regions respectively to obtain a plurality of remote sensing images; A splicing unit connected to the remote sensing shooting unit for splicing the plurality of remote sensing images according to the same size ratio to obtain the overall spliced image of the area to be detected on the outer surface of the containment vessel; A grayscale processing unit, connected to the splicing unit, for performing grayscale processing on the overall spliced image to obtain the image to be detected on the outer surface of the containment vessel.

Images