X-ray image-based aero-engine part defect detection method

Abstract

The invention relates to the field of defect detection, in particular to an aero-engine part defect detection method based on an X-ray image, and the method comprises the steps: obtaining the X-ray image of a part, and segmenting the X-ray image into a plurality of sub-regions; and calculating the optimal gamma value of each sub-region, performing gamma transformation on each sub-region by using the optimal gamma value to obtain an optimal X-ray image, and inputting the optimal X-ray image into a preset detection model to obtain the type of the part defect. According to the invention, the accuracy of part defect detection results can be improved.

Description

Technical Field

[0001] The invention relates to the field of defect detection, and in particular to an aerospace engine component defect detection method based on X-ray images. Background Art

[0002] As the core component of an aircraft, the performance and reliability of an aircraft engine are directly related to flight safety. In the production process of aircraft engine parts, various defects such as cracks, pores, inclusions, etc. will occur due to various factors such as materials and processes. Defective parts will affect the performance and safety of the engine. Therefore, it is necessary to perform defect detection on aircraft engine parts during the production process. In the production process of aircraft engine parts, non-destructive testing technology is used for defect detection. Common non-destructive testing methods include ultrasonic testing, eddy current testing, X-ray testing, magnetic particle testing, penetration testing, etc. These methods can detect internal and surface defects without destroying the parts.

[0003] A Chinese patent document with announcement number CN116051542B discloses a defect detection method and a defect detection device, the defect detection method comprising: scanning a training sample to obtain a first image, wherein the training sample has a welding defect; denoising the first image to obtain a second image; generating a three-dimensional point cloud image and a brightness image of the training sample based on the second image; training and identifying the welding defect based on the three-dimensional point cloud image and the brightness image; generating a defect recognition model after the training is completed; and using the defect recognition model to perform welding defect detection on the object to be detected.

[0004] When using the detection model to detect aircraft engine parts, it is necessary to enhance the image of the parts to show the texture features of the surface of the parts and improve the accuracy of the detection results. The commonly used method for image enhancement is gamma transformation. However, due to the different grayscale performances of different edge features in the image, especially when the background brightness of each part of the image is different, the same gamma coefficient cannot simultaneously meet the enhancement effect of different edges and details in the image. And the more obvious the edge feature is, the greater the degree of contrast enhancement needs to be. Therefore, the use of traditional gamma transformation cannot meet the corresponding requirements, resulting in poor enhancement of X-ray images, thereby affecting the accuracy of defect detection on aircraft engine parts. Summary of the invention

[0005] In order to solve the problem of low accuracy of defect detection caused by poor enhancement effect on X-ray images, the present invention provides an aircraft engine component defect detection method based on X-ray images.

[0006] The present invention provides an aerospace engine parts defect detection method based on X-ray images, which adopts the following technical solutions: The method for detecting defects of aircraft engine parts based on X-ray images comprises the following steps: Acquire an X-ray image of a component, and divide the X-ray image into a plurality of sub-regions; Calculate the optimal gamma value of each sub-region, use the optimal gamma value to perform gamma transformation on each sub-region to obtain the optimal X-ray image, and input the optimal X-ray image into the preset detection model to obtain the type of component defects; wherein, the calculation method of the optimal gamma value of the sub-region is: Perform binary classification on the pixels in the sub-region to obtain a first category and a second category; use the longest boundary line between the pixels in the first category and the pixels in the second category as a segmentation line; Calculate the rationality of the binary classification, which represents the accuracy of the binary classification results; calculate the consistency and continuity of the gradient direction of the segmentation line in the sub-region, and continuity and consistency are positively correlated; The expression for the optimal gamma value is: ; In the formula, γ represents the optimal gamma value of the sub-region, W represents the rationality of the first category and the second category in the sub-region, Y represents the gradient direction consistency of the segmentation line in the sub-region, and G represents the continuity of the segmentation line in the sub-region. Represents the mean grayscale value of pixels in the sub-region. represents the median grayscale value of the pixel in the sub-region, p represents the preset first hyperparameter, Represents the preset second hyperparameter.

[0007] By calculating the optimal gamma value of each sub-region of the X-ray image and using the optimal gamma value to perform gamma transformation on the corresponding sub-region, the feature contrast of the X-ray image is improved, and the accuracy of defect detection is further improved.

[0008] Preferably, the method of performing binary classification on the pixels in the sub-region to obtain the first category and the second category is: Construct a feature vector for each pixel, and the expression of the feature vector is: ; in, Indicates the first The feature vector of pixels, Indicates the first The gray value of a pixel, Represents the maximum gray value in the sub-region, Represents the minimum gray value in the sub-region, Indicates The number of columns where the pixel point is located in the sub-region, Indicates The number of rows where the pixel point is located in the sub-region; The feature vectors of the pixels are clustered to obtain the first category and the second category.

[0009] The feature vector is constructed by the pixel position and grayscale features, and then the pixel points are classified using the feature vector, which improves the accuracy of the classification results.

[0010] Preferably, the calculation method for the rationality of the two-class classification is:

[0011] In the formula, Indicates the rationality of dividing the sub-areas into the first and second categories, It represents the distance between the two cluster centers of the first and second categories. Indicates the first category The distance from a pixel to the center point of the class, Represents the distance from the Lth pixel in the second category to the center point within the category, Represents the maximum gray value in the sub-region, represents the minimum gray value in the sub-region, exp represents the exponential function with e as the base, and m represents the number of pixels in the first category.

[0012] The above formula can be used to quantify the accuracy of the classification results, and the rationality can be used to preliminarily understand whether the corresponding sub-region is a flat area.

[0013] Preferably, the gradient direction consistency calculation method of the segmentation line is:

[0014] In the formula, Indicates the consistency of the gradient direction of the segmentation line in the sub-region, Indicates the azimuth angle of the dividing line next to the k-th pixel on one side of the dividing line. Indicates the gradient direction of the k-th pixel on one side of the dividing line, represents the gradient direction of the k+1th pixel on one side of the dividing line, and t represents the total number of pixels on one side of the dividing line.

[0015] The consistency of the gradient direction of the segmentation line is calculated in multiple dimensions, which improves the accuracy of the calculation result. The correctness of the segmentation line can be understood through the consistency of the gradient direction of the segmentation line.

[0016] Preferably, the calculation method of the continuity of the dividing line in the sub-region is: The sub-region to be calculated for continuity is taken as the target region, the continuity of the segmentation line in the sub-region adjacent to the target region is calculated, and the maximum value among the continuity of the segmentation line in the adjacent sub-region is taken as the continuity of the segmentation line in the target region.

[0017] Preferably, the expression for the continuity of the dividing lines in adjacent sub-regions is:

[0018] In the formula, represents the continuity of the segmentation line in the u-th sub-region adjacent to the target region, represents the gradient direction consistency of the segmentation line in the u-th sub-region adjacent to the target region, Y represents the gradient direction consistency of the segmentation line in the target region, represents the distance between the endpoint of the segmentation line in the target area and the endpoint of the segmentation line in the adjacent u-th sub-area, Indicates the angle between the target region and the extended line of the endpoint of the adjacent region dividing line.

[0019] Preferably, the expression for the continuity of the dividing lines in adjacent sub-regions is:

[0020] In the formula, Indicates the first The continuity of the dividing line in each sub-area, represents the gradient direction consistency of the segmentation line in the u-th sub-region adjacent to the target region, Y represents the gradient direction consistency of the segmentation line in the target region, Indicates the endpoints of the segmentation line in the target area and the adjacent The distance between the endpoints of the dividing line in each sub-area.

[0021] Preferably, before inputting the best X-ray image into the preset detection model, the method also includes a step of performing mean filtering on the pixel points at the boundary of the sub-region.

[0022] Mean filtering of the pixel points at the boundary of the sub-region can avoid the grayscale value mutation of the boundary pixel points of each sub-region, make the grayscale connection between the sub-regions more natural, and avoid the block effect of the grayscale value.

[0023] The present invention has the following technical effects: By calculating the optimal gamma value of each sub-region in the X-ray image and using the optimal gamma value to perform gamma transformation on the corresponding sub-region, the feature contrast of the X-ray image is improved compared with the traditional gamma transformation, and the accuracy of defect detection is further improved. BRIEF DESCRIPTION OF THE DRAWINGS

[0024] By reading the detailed description below with reference to the accompanying drawings, the above and other purposes, features and advantages of the exemplary embodiments of the present invention will become readily understood. In the accompanying drawings, several embodiments of the present invention are shown in an exemplary and non-restrictive manner, and the same or corresponding reference numerals represent the same or corresponding parts.

[0025] Figure 1 The present invention is a flow chart of an aircraft engine parts defect detection method based on X-ray images.

[0026] Figure 2 It is a schematic diagram of the dividing line of the present invention.

[0027] Figure 3 2 is another schematic diagram of the dividing line of the present invention. DETAILED DESCRIPTION

[0028] The following will be combined with the drawings in the embodiments of the present invention to clearly and completely describe the technical solutions in the embodiments of the present invention. Obviously, the described embodiments are part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative work are within the scope of protection of the present invention.

[0029] It should be understood that when the terms "first", "second", etc. are used in the claims, descriptions, and drawings of the present invention, they are only used to distinguish different objects, rather than to describe a specific order. The terms "include" and "comprise" used in the description and claims of the present invention indicate the presence of the described features, wholes, steps, operations, elements, and / or components, but do not exclude the presence or addition of one or more other features, wholes, steps, operations, elements, components, and / or their collections.

[0030] The embodiment of the present invention discloses a method for detecting defects of aircraft engine parts based on X-ray images. Figure 1 , including the following steps: S1: Acquire an X-ray image of a component and divide the X-ray image into multiple sub-areas.

[0031] Obtain an X-ray image of an aircraft engine component. The X-ray image is a grayscale area. Divide the X-ray image into multiple 5×5 sub-areas, that is, each sub-area contains 25 pixels.

[0032] S2: Classify the pixels in the sub-region into a first category and a second category, and use the boundary between the pixels in the first category and the pixels in the second category as a segmentation line.

[0033] In one embodiment, in each sub-region, a feature vector is constructed for each pixel; the expression of the feature vector is: ;in, Indicates the first The feature vector of pixels, Indicates the first The gray value of a pixel, Represents the maximum gray value in the sub-region, Represents the minimum gray value in the sub-region, Indicates The number of columns where the pixel point is located in the sub-region, Indicates The number of rows where the pixel points are located in the sub-region; the feature vectors of the pixel points are clustered using the k-means clustering algorithm to obtain the first category and the second category. It can be understood that the process of clustering the feature vectors using the k-means clustering algorithm is the process of dividing the pixel points in the sub-region into two categories, wherein the first category includes multiple pixel points, and the second category includes multiple pixel points.

[0034] In each sub-area, the pixels are divided into two categories, wherein the pixels in the first category and the pixels in the second category have multiple boundary lines, and the longest boundary line is used as the dividing line. It should be noted that the dividing line can be a closed line or an open line.

[0035] For example, in Figure 2 In the figure, the white area is the pixel points of the first category, the shaded area is the pixel points of the second category, there are dividing line 1 and dividing line 2, and both dividing line 1 and dividing line 2 are closed lines, wherein the length of dividing line 1 is greater than the length of dividing line 2, therefore, dividing line 1 is used as the dividing line of the sub-area.

[0036] exist Figure 3 In the figure, the white area is the pixel points of the first category, the shadow area is the pixel points of the second category, there are a dividing line 1 and a dividing line 2, the dividing line 1 is a closed line, and the dividing line 2 is an open line, wherein the length of the dividing line 1 is greater than the length of the dividing line 2, therefore, the dividing line 1 is used as the dividing line of the sub-area. It can be understood that the dividing line is composed of multiple sub-dividing lines.

[0037] In one embodiment, in each sub-region, the pixels are classified into two categories using the Otsu threshold segmentation method.

[0038] S3: Calculate the rationality of the binary classification. The rationality represents the accuracy of the binary classification results.

[0039] The calculation method for the rationality of the two classifications is:

[0040] In the formula, Indicates the rationality of dividing the sub-areas into the first and second categories, It represents the distance between the two cluster centers of the first and second categories. Indicates the first category The distance from a pixel to the center point of the class, Represents the distance from the Lth pixel in the second category to the center point within the category, Represents the maximum gray value in the sub-region, represents the minimum gray value in the sub-region, exp represents the exponential function with e as the base, and m represents the number of pixels in the first category.

[0041] It represents the difference between the maximum grayscale value and the minimum grayscale value in the sub-region. The larger its value is, the larger the grayscale value span in the target area block is, and the greater the possibility that there is an edge in the corresponding sub-region is, that is, the corresponding sub-region may be at the boundary of two grayscale changes (for example, the boundary between the foreground and the background, the boundary of the defect, etc.), and the more reasonable it is to divide the pixels in the corresponding sub-region into two categories; the smaller its value is, the smaller the grayscale value span in the corresponding sub-region is, the greater the possibility that the corresponding sub-region is a flat area, and the less reasonable it is to divide it into two categories.

[0042] It represents the distance between the cluster centers of two categories. The larger the value, the more reasonable it is to divide the corresponding sub-region into two categories; when The smaller the value, the more reasonable it is to divide the corresponding sub-area into two categories. The smaller.

[0043] It represents the average distance from the pixels in two categories to the center point within the category, reflecting the density of elements within the category. The larger its value is, the larger the average distance from the pixels in the category to the center point within the category is, indicating that the category density is smaller, thus The smaller it is, the less reasonable it is to divide the pixels in the corresponding sub-region into two categories; when The smaller the value, the smaller the average distance from the pixel point in the category to the center point in the category, which means the category density is greater. The larger the value, the more reasonable it is to divide the target area blocks into two categories. The bigger it is.

[0044] S4: Calculate the gradient direction consistency of the segmentation line in the sub-region.

[0045] The gradient direction consistency calculation method of the segmentation line is:

[0046] In the formula, Indicates the consistency of the gradient direction of the segmentation line in the sub-region, Indicates the azimuth angle of the dividing line next to the k-th pixel on one side of the dividing line. The azimuth angles are divided into up, down, left, and right, and the corresponding angles are 90 degrees, -90 degrees, 180 degrees, and 0 degrees, respectively. Indicates the gradient direction of the k-th pixel on one side of the dividing line, represents the gradient direction of the k+1th pixel on one side of the dividing line, and t represents the total number of pixels on one side of the dividing line. It should be noted that one side of the dividing line refers to the inner side or the outer side of the dividing line. In the actual calculation process, one side is selected. The Sobel operator is used to calculate the gradient direction of the pixel points in the sub-region.

[0047] when When the value is larger, it means that the gradient direction of the kth pixel is less consistent with the orientation of the dividing line, which means that the gradient direction of the pixel is less perpendicular to the dividing line, which means that the accuracy of the dividing line is smaller. The smaller. The smaller the time, the more The more consistent the gradient direction of a pixel is with the orientation of the dividing line, the more perpendicular the gradient direction of the pixel is to the dividing line, which means the correctness of the dividing line is greater. The bigger.

[0048] when The larger the value, the Pixels and The gradient directions of the pixels are quite different, which indicates that The gradient direction consistency at each pixel is small; when The smaller the time, the more Pixels and The gradient directions of the pixels are slightly different, indicating that The gradient direction at each pixel is more consistent.

[0049] S5: Calculate the continuity of the dividing line within the sub-region. Continuity is positively correlated with consistency.

[0050] The sub-region to be calculated for continuity is taken as the target region, the continuity of the segmentation line in the sub-region adjacent to the target region is calculated, and the maximum value among the continuity of the segmentation line in the adjacent sub-region is taken as the continuity of the segmentation line in the target region.

[0051] In one embodiment, the expression for the continuity of the segmentation lines in adjacent sub-regions is:

[0052] In the formula, represents the continuity of the segmentation line in the u-th sub-region adjacent to the target region, represents the gradient direction consistency of the segmentation line in the u-th sub-region adjacent to the target region, Y represents the gradient direction consistency of the segmentation line in the target region, represents the distance between the endpoint of the segmentation line in the target area and the endpoint of the segmentation line in the adjacent u-th sub-area, β represents the angle between the extension line of the endpoint of the segmentation line in the target area and the adjacent area, and the extension line and the local segmentation line where the endpoint is located are located in the same straight line direction. It should be noted here that in an open segmentation line, an endpoint close to another sub-area is selected. For example, for sub-area a and sub-area b, the segmentation line in sub-area a has endpoints 1 and 2, and sub-area b has endpoints 3 and 4, where endpoint 1 is close to sub-area b, and endpoint 3 is close to sub-area a, then the angle between the extension line of endpoint 1 and the extension line of endpoint 3 is β. In a closed segmentation line, a point on the sub-segmentation line that is closer to another sub-area is selected as the endpoint.

[0053] Therefore, the expression for the continuity of the segmentation line in the target area is: , where G represents the continuity of the segmentation line in the target area, Indicates the continuity of the segmentation line in the nth sub-region adjacent to the target region.

[0054] In one embodiment, the expression for the continuity of the segmentation lines in adjacent sub-regions is:

[0055] In the formula, represents the continuity of the segmentation line in the u-th sub-region adjacent to the target region, represents the gradient direction consistency of the segmentation line in the u-th sub-region adjacent to the target region, Y represents the gradient direction consistency of the segmentation line in the target region, Represents the distance between the endpoint of the segmentation line in the target region and the endpoint of the segmentation line in the adjacent u-th sub-region.

[0056] S6: Calculate the optimal gamma value of each sub-region, and use the optimal gamma value to perform gamma transformation on each sub-region to obtain an optimal X-ray image.

[0057] The expression for the optimal gamma value is: ; In the formula, γ represents the optimal gamma value of the sub-region, W represents the rationality of the first category and the second category in the sub-region, Y represents the gradient direction consistency of the segmentation line in the sub-region, and G represents the continuity of the segmentation line in the sub-region. Represents the mean grayscale value of pixels in the sub-region. represents the median of the grayscale values of the pixel points in the sub-region, p represents the preset first hyperparameter, represents the preset second hyperparameter. The values of the first hyperparameter and the second hyperparameter can be set manually according to the actual situation. Exemplarily, the value of p is 0.2, and the value of

[0058] is 0.4. When W < p or G < σ, it indicates that the corresponding sub-region is very likely to be a flat region or a flat region containing noise. At this time, the gamma coefficient is 1, which is equivalent to not changing the contrast of the corresponding sub-region.

[0059] When W ≥ p and G ≥ σ, it indicates that there are regular grayscale change characteristics in the corresponding sub-region. In addition to the structural edges of the components themselves, there may be defects or abnormal grayscale changes. Therefore, it is necessary to enhance the contrast of the target region block to make the grayscale changes that are not obvious in the original X-ray image more obvious. At this time, if , it indicates that the sub-region belongs to the region with a larger grayscale value, that is, the brighter region. At this time, the grayscale change characteristics may not be obvious due to reasons such as exposure. Therefore, the gamma coefficient is set. At this time, the gamma coefficient γ is greater than 1, which can increase the contrast of the brighter sub-region. And when the grayscale change characteristics of the sub-region are more obvious, that is, the larger W + Y + G is, the gamma coefficient is larger, so the degree of enhancement of the contrast of the corresponding sub-region is greater, and the grayscale change characteristics of the sub-region can be enhanced to a greater extent. If , it indicates that the sub-region belongs to the region with a smaller grayscale value, that is, the darker region. At this time, the grayscale change characteristics may not be obvious due to reasons such as lack of light. Therefore, the gamma coefficient is set. At this time, the gamma coefficient is less than 1, which can increase the contrast of the darker sub-region. And when the grayscale change characteristics of the corresponding sub-region are more obvious, that is, the larger W + Y + G is, the gamma coefficient is smaller, and the degree of enhancement of the contrast of the sub-region is greater, and the grayscale change characteristics of the target region block can be enhanced to a greater extent.

[0060] Each sub-region corresponds to an optimal gamma value. The grayscale values of each pixel point in the corresponding sub-region are subjected to gamma transformation using the calculated optimal gamma value. The formula is as follows:

[0061] where, represents the grayscale value after gamma transformation of the th pixel point in the sub-region, represents the original grayscale value of the th pixel point in the sub-region before gamma transformation, is the gamma coefficient of the target area block. The floor symbol.

[0062] The entire X-ray image after gamma change is taken as the optimal X-ray image. In order to avoid sudden changes in grayscale values ​​at the boundary pixels of each sub-region, a 3×3 window is used to perform mean filtering on the pixels at the boundary of the sub-region, so that the grayscale connection between the sub-regions is more natural and the block effect of the grayscale value is avoided.

[0063] S7: Input the best X-ray image into a preset detection model to obtain the type of component defects.

[0064] The method for obtaining the detection model is to construct a yolov8 algorithm model, obtain a training set, and use the training set to train the yolov8 algorithm model to obtain a detection model. The training of the yolov8 algorithm model is an existing technology, and the specific steps are not repeated here.

[0065] The best X-ray image is input into a preset detection model to obtain the types of component defects, where the types of component defects are divided into pores, cracks, and impurities.

[0066] Although this specification has shown and described a number of embodiments of the present invention, it will be apparent to those skilled in the art that such embodiments are provided by way of example only. Those skilled in the art will conceive of many modifications, changes and alternatives without departing from the ideas and spirit of the present invention. It should be understood that in the practice of the present invention, various alternatives to the embodiments of the present invention described herein may be employed.

[0067] The above are all preferred embodiments of the present invention, and are not intended to limit the protection scope of the present invention. Therefore, any equivalent changes made based on the structure, shape, and principle of the present invention should be included in the protection scope of the present invention.

Claims

1. A method for detecting defects in aircraft engine parts based on X-ray images, characterized in that: The following steps are involved: Acquire an X-ray image of a component, and divide the X-ray image into a plurality of sub-regions; Calculate the optimal gamma value of each sub-region, use the optimal gamma value to perform gamma transformation on each sub-region to obtain the optimal X-ray image, and input the optimal X-ray image into the preset detection model to obtain the type of component defects; wherein, the calculation method of the optimal gamma value of the sub-region is: Perform binary classification on the pixels in the sub-region to obtain a first category and a second category; use the longest boundary line between the pixels in the first category and the pixels in the second category as a segmentation line; Calculate the rationality of the binary classification, which represents the accuracy of the binary classification results; calculate the consistency and continuity of the gradient direction of the segmentation line in the sub-region, and continuity and consistency are positively correlated; The expression for the optimal gamma value is: ； In the formula, γ represents the optimal gamma value of the sub-region, W represents the rationality of the first category and the second category in the sub-region, Y represents the gradient direction consistency of the segmentation line in the sub-region, and G represents the continuity of the segmentation line in the sub-region. Represents the mean grayscale value of pixels in the sub-region. represents the median grayscale value of the pixel in the sub-region, p represents the preset first hyperparameter, Represents the preset second hyperparameter.

2. The method for detecting defects of aircraft engine parts based on X-ray images according to claim 1, characterized in that: The method of binary classification of pixels in the sub-region to obtain the first category and the second category is: Construct a feature vector for each pixel, and the expression of the feature vector is: ； in, Indicates the first The feature vector of pixels, Indicates the first The gray value of a pixel, Represents the maximum gray value in the sub-region, Represents the minimum gray value in the sub-region, Indicates The number of columns where the pixel point is located in the sub-region, Indicates The number of rows where the pixel point is located in the sub-region; The feature vectors of the pixels are clustered to obtain the first category and the second category.

3. The method for detecting defects of aircraft engine parts based on X-ray images according to claim 2, characterized in that: The calculation method for the rationality of the two classifications is: In the formula, Indicates the rationality of dividing the sub-areas into the first and second categories, It represents the distance between the two cluster centers of the first and second categories. Indicates the first category The distance from a pixel to the center point of the class, Represents the distance from the Lth pixel in the second category to the center point within the category, Represents the maximum gray value in the sub-region, represents the minimum gray value in the sub-region, exp represents the exponential function with e as the base, and m represents the number of pixels in the first category.

4. The method for detecting defects of aircraft engine parts based on X-ray images according to claim 1, characterized in that: The gradient direction consistency calculation method of the segmentation line is: In the formula, Indicates the consistency of the gradient direction of the segmentation line in the sub-region, Indicates the azimuth angle of the dividing line next to the k-th pixel on one side of the dividing line. Indicates the gradient direction of the k-th pixel on one side of the dividing line, represents the gradient direction of the k+1th pixel on one side of the dividing line, and t represents the total number of pixels on one side of the dividing line.

5. The method for detecting defects of aircraft engine parts based on X-ray images according to claim 1, characterized in that: The continuity of the dividing line within a sub-region is calculated as: The sub-region to be calculated for continuity is taken as the target region, the continuity of the segmentation line in the sub-region adjacent to the target region is calculated, and the maximum value among the continuity of the segmentation line in the adjacent sub-region is taken as the continuity of the segmentation line in the target region.

6. The method for detecting defects of aircraft engine parts based on X-ray images according to claim 5, characterized in that: The expression for the continuity of the dividing lines in adjacent sub-regions is: In the formula, Indicates the first The continuity of the dividing line in each sub-area, represents the gradient direction consistency of the segmentation line in the u-th sub-region adjacent to the target region, Y represents the gradient direction consistency of the segmentation line in the target region, Indicates the endpoints of the segmentation line in the target area and the adjacent The distance between the endpoints of the dividing line in each sub-area, Indicates the angle between the target region and the extended line of the endpoint of the adjacent region dividing line.

7. The method for detecting defects of aircraft engine parts based on X-ray images according to claim 5, characterized in that: The expression for the continuity of the dividing lines in adjacent sub-regions is: In the formula, represents the continuity of the segmentation line in the u-th sub-region adjacent to the target region, represents the gradient direction consistency of the segmentation line in the u-th sub-region adjacent to the target region, Y represents the gradient direction consistency of the segmentation line in the target region, Represents the distance between the endpoint of the segmentation line in the target region and the endpoint of the segmentation line in the adjacent u-th sub-region.

8. The method for detecting defects of aircraft engine parts based on X-ray images according to claim 1, characterized in that: Before inputting the best X-ray image into the preset detection model, the method also includes the step of performing mean filtering on the pixel points at the boundary of the sub-region.

Images