A method and system for detecting surface defects of a turbine assembly of an air cycle machine

Abstract

The present invention discloses a method and system for detecting surface defects of turbine components of an air cycle machine, comprising: building a detection basis through image preprocessing and edge extraction, enhancing defect feature contrast in combination with grayscale value and gradient amplitude correction, realizing refined regional division using a region growing algorithm, quantifying defect judgment criteria with a scoring coefficient, and optimizing detection results through overlapping region screening, thereby effectively improving the accuracy and reliability of defect detection, being able to accurately identify minor defects and avoid misjudgment, and at the same time, the automated processing flow significantly improving detection efficiency, providing efficient and intelligent technical support for turbine component quality control.

Description

Technical Field

[0001] The present invention belongs to the technical field of image processing and relates to a method and system for detecting surface defects of a turbine component of an air cycle machine. Background Art

[0002] During long-term use, the turbine blades of air cycle machines may develop surface defects such as cracks, wear, and corrosion due to high temperature, high pressure, and complex airflow. These defects will not only affect the performance of the turbine and reduce work efficiency, but may also cause damage to the equipment. In order to improve the inspection accuracy and efficiency of turbine blades, it is particularly important to adopt effective surface defect detection technology. Traditional methods for detecting surface defects of turbine blades, such as manual inspection or traditional machine vision technology, often rely on manual experience or relatively complex algorithms, with low detection efficiency, poor accuracy, and susceptibility to environmental interference. In addition, the structural features of turbine blades are too obvious, and the defect features on the surface of turbine blades, such as grayscale values ​​and gradient amplitudes, are too close to the normal area, which makes the defects difficult to detect during the inspection process. Therefore, there is an urgent need for a method that can efficiently and accurately detect surface defects of turbine blades. Summary of the Invention

[0003] The purpose of the present invention is to solve the problems in the prior art of low efficiency, poor accuracy and susceptibility to environmental interference in traditional turbine blade surface defect detection, and to provide a method and system for detecting surface defects of turbine components of an air cycle machine.

[0004] To achieve the above-mentioned object, the present invention adopts the following technical solutions: a method for detecting surface defects of a turbine assembly of an air cycle machine, comprising: acquiring an image of a turbine blade to be inspected and preprocessing the acquired image; extracting edge information of the turbine blade based on a Canny algorithm to obtain an edge image of the turbine blade and grayscale values ​​of pixels in the edge image, and obtaining gradient amplitudes of all pixels in the edge image using an edge algorithm; correcting the grayscale values ​​and gradient amplitudes of the pixels in the edge image; treating the area enclosed by closed edge pixels in the edge image as a plurality of independent large areas; for any large area, dividing the area based on the gradient amplitude correction values ​​and grayscale correction values ​​of pixels in the area other than the edge pixels of the large area to obtain a plurality of small areas; repeating this step to divide all large areas; obtaining a scoring coefficient for each small area based on the pixel information in the divided small areas, and then determining whether the current small area is a defective area; if there is overlap between adjacent defective areas, screening the true defective area based on the overlapping area between the adjacent defective areas.

[0005] A further improvement of the present invention is that: further, the collected image is preprocessed, specifically: the collected turbine blade image is grayscaled and denoised, and the turbofan is segmented from the image using image segmentation technology to obtain a preprocessed turbine blade grayscale image.

[0006] Furthermore, the grayscale values ​​of the pixels in the edge image are corrected, specifically: in, Indicates that it is located Grayscale correction value of the pixel at Indicates that it is located The original gray value of the pixel at Indicates the maximum value of the pixel grayscale value in the original grayscale image of the turbine blade; Indicates the minimum grayscale value of the pixel in the original grayscale image of the turbine blade;

[0007] Correct the gradient amplitude of the pixel points in the edge image, specifically: in, Indicates that it is located The gradient amplitude correction value of the pixel point at ; Indicates that among the remaining pixels in the original edge image that have been determined as edge pixels, The original gradient amplitude of the pixel at ; Indicates the maximum value of the original gradient amplitude among the remaining pixels except those that have been determined as edge pixels in the original edge image; Indicates the minimum value of the original gradient amplitude among the remaining pixels in the original edge image except those that have been determined as edge pixels.

[0008] Furthermore, the area enclosed by the closed edge pixels in the edge image is regarded as several independent large areas. For any large area, the area is divided based on the gradient amplitude correction value and the pixel grayscale correction value of the pixel points other than the edge pixels of the large area in the area to obtain several small areas. This step is repeated to divide all the large areas. Specifically, based on the gradient amplitude correction value and the pixel grayscale correction value of the pixel points other than the edge pixels of the large area in the area, the judgment coefficient of the pixel points other than the edge pixels of the large area in the divided large area is obtained; the pixel point with the largest judgment coefficient is regarded as the growth center pixel point. Then, a growing window is constructed; based on the pixel points between the growing windows, the defect coefficient of the growing center pixel point in the divided large area is obtained; any growing window is selected as the central growing window, and the central growing window grows toward the adjacent growing window, and based on the ratio of the defect coefficient between the central growing window and the adjacent growing window, it is judged whether the adjacent growing window and the central growing window are in the same area; if so, the adjacent growing window is selected as the center, and growth is continued with other growing windows until the growing window is filled and cannot grow; if not, growth is stopped; the above steps are repeated until all pixel points in the large area are divided into small areas, and the small area division operation is completed.

[0009] Furthermore, the judgment coefficients of the pixels in the divided large area except the edge pixels of the large area are obtained, specifically: in, Indicates that the pixels in the divided large area are located outside the edge of the large area. Judgment coefficient of position pixel point; Indicates that the pixels in the divided large area are located outside the edge of the large area. The gradient amplitude correction value of the position pixel point; Indicates that the pixels in the divided large area are excluding the edge pixels of the large area and the pixels located in The gradient amplitude correction value of the pixel points other than the position pixel point; Indicates that the pixels in the divided large area are excluding the edge pixels of the large area and the pixels located in The number of pixels outside the position pixel; Indicates that the pixels in the divided large area are located outside the edge of the large area. Grayscale correction value of the pixel at the position; Indicates that the pixels in the divided large area are excluding the edge pixels of the large area and the pixels located in Grayscale correction value of pixels other than the position pixel.

[0010] Furthermore, the pixel with the largest judgment coefficient is regarded as the growth center pixel, and then a growth window is constructed. Based on the pixels between the growth windows, the defect coefficient of the growth center pixel in the divided large area is obtained. Specifically, the pixel with the largest judgment coefficient is regarded as the growth center pixel, and a window of size n*n is established, with the center of the window as the growth center pixel, and the growth is performed according to the following rules: in, Indicates the defect coefficient of the growth center pixel in the divided large area; Represents the average value of the gradient amplitude correction value of all pixels in the window; Indicates the variance of the gradient amplitude correction values ​​of all pixels in the window; Grayscale correction value of growth center pixel in the divided large area; Indicates the average grayscale correction value of the remaining pixels in the window except the center pixel of the window.

[0011] Furthermore, based on the ratio of the defect coefficients between the central growth window and the adjacent growth window, it is determined whether the adjacent growth window and the central growth window are in the same area. Specifically, when the ratio of the defect coefficients between the central growth window and the adjacent growth window is greater than the set threshold, the adjacent growth window and the central growth window are considered to be in the same area; wherein, when the defect coefficients between the central growth window and the adjacent growth window are compared, the one with the larger defect coefficient is used as the denominator.

[0012] Furthermore, based on the pixel information in the divided small areas, the scoring coefficient of each small area is obtained, specifically: in, Indicates the The scoring coefficient of a small area; Indicates the The average value of the defect coefficient of each pixel in a small area; Indicates the minimum value of the mean defect coefficient of all small areas; Indicates the The number of pixels in a small area; the judgment of whether the current small area is a defect area is specifically: based on the obtained scoring coefficient of each small area , get the average score coefficient of the whole picture ; If Rating coefficient of a small area Greater than the average rating coefficient of the entire map , then the corresponding small area is marked as a defect area.

[0013] Furthermore, based on the overlapping area between adjacent defective areas, the real defective areas are screened out, specifically: the ratio of the overlapping area between adjacent defective areas to the total area of ​​the corresponding two adjacent areas is obtained; and it is determined whether the obtained ratio is greater than a preset threshold. If it exceeds, it means that there is a serious overlap between the two areas, and the scoring coefficient is retained at this time. The high defect areas are selected and the rest of the areas are eliminated; the real defect areas are obtained by screening.

[0014] A surface defect detection system for a turbine assembly of an air cycle machine includes: a preprocessing module, which collects an image of a turbine blade to be detected and preprocesses the collected image; an acquisition module, which extracts edge information of the turbine blade based on a Canny algorithm, obtains an edge image of the turbine blade and the grayscale value of the pixel points in the edge image, and obtains the gradient amplitude of all pixel points in the edge image through an edge algorithm; a correction module, which corrects the grayscale value of the pixel points and the gradient amplitude of the pixel points in the edge image respectively; a segmentation module, which divides the edge image into the pixels enclosed by the closed edge pixels. The formed areas are regarded as several independent large areas. For any large area, the area is divided based on the gradient amplitude correction value and the pixel grayscale correction value of the pixel points in the area except the edge pixels of the large area to obtain several small areas; and this step is repeated to divide all the large areas; a judgment module is used, which obtains the scoring coefficient of each small area based on the pixel information in the divided small areas, and then determines whether the current small area is a defect area; a screening module is used, which screens out the real defect area based on the overlapping area between the adjacent defect areas if there is overlap between the adjacent defect areas.

[0015] Compared with the existing technology, the present invention has the following beneficial effects: the present invention constructs a detection basis through image preprocessing and edge extraction, combines grayscale value and gradient amplitude correction to enhance the contrast of defect features, uses region growing algorithm to achieve refined region division, quantifies defect judgment criteria with scoring coefficients, and optimizes detection results through overlapping area screening, effectively improving the accuracy and reliability of defect detection, being able to accurately identify minor defects and avoid misjudgment, and at the same time, the automated processing flow significantly improves detection efficiency, providing efficient and intelligent technical support for turbine component quality control. BRIEF DESCRIPTION OF THE DRAWINGS

[0016] In order to more clearly illustrate the technical solutions of the embodiments of the present invention, the following briefly introduces the drawings required for use in the embodiments. It should be understood that the following drawings only illustrate certain embodiments of the present invention and therefore should not be regarded as limiting the scope. For ordinary technicians in this field, other relevant drawings can be obtained based on these drawings without paying any creative work.

[0017] Figure 1 Schematic diagram of a flow chart of a method for detecting surface defects of a turbine assembly of an air cycle machine according to the present invention; Figure 2 Schematic diagram of the structure of a surface defect detection system for a turbine assembly of an air cycle machine according to the present invention; Figure 3 This is the image of the turbine blade edge; Figure 4 A growth diagram. DETAILED DESCRIPTION

[0018] To make the objectives, technical solutions, and advantages of the embodiments of the present invention more clear, the technical solutions of the embodiments of the present invention will be clearly and completely described below in conjunction with the accompanying drawings of the embodiments of the present invention. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. Generally, the components of the embodiments of the present invention described and shown in the drawings herein can be arranged and designed in various different configurations.

[0019] Therefore, the following detailed description of the embodiments of the present invention provided in the accompanying drawings is not intended to limit the scope of the invention as claimed, but rather merely represents selected embodiments of the present invention. All other embodiments derived by persons of ordinary skill in the art based on the embodiments of the present invention without creative effort shall fall within the scope of protection of the present invention.

[0020] It should be noted that similar reference numerals and letters denote similar items in the following drawings, and therefore, once an item is defined in one drawing, it does not need to be further defined or explained in subsequent drawings.

[0021] In the description of the embodiments of the present invention, it should be noted that if the terms "upper," "lower," "horizontal," "inner," etc. appear, the orientation or positional relationship indicated is based on the orientation or positional relationship shown in the accompanying drawings, or the orientation or positional relationship in which the inventive product is typically placed when in use. These terms are merely for the convenience of describing the present invention and simplifying the description, and do not indicate or imply that the device or component referred to must have a specific orientation, be constructed, or operate in a specific orientation. Therefore, they should not be construed as limitations on the present invention. In addition, the terms "first," "second," etc. are used only to distinguish descriptions and should not be construed as indicating or implying relative importance.

[0022] In addition, if the term "horizontal" appears, it does not mean that the component must be absolutely horizontal, but can be slightly tilted. For example, "horizontal" only means that its direction is more horizontal than "vertical", and does not mean that the structure must be completely horizontal, but can be slightly tilted.

[0023] In the description of the embodiments of the present invention, it should be noted that, unless otherwise expressly specified or limited, the terms "disposed," "installed," "connected," and "connected" should be understood in a broad sense. For example, they can refer to fixed connections, detachable connections, or integral connections; they can refer to mechanical connections or electrical connections; they can refer to direct connections or indirect connections through an intermediate medium; and they can refer to internal connections between two components. Those skilled in the art will understand the specific meanings of the above terms in the present invention based on specific circumstances.

[0024] The present invention will be described in further detail below with reference to the accompanying drawings: Figure 1 The present invention discloses a method for detecting surface defects of a turbine assembly of an air cycle machine, comprising: S101, acquiring an image of a turbine blade to be inspected, and preprocessing the acquired image; performing grayscale denoising processing on the acquired turbine blade image, and using image segmentation technology to segment the turbofan from the image to obtain a preprocessed grayscale image of the turbine blade.

[0025] S102, extracting edge information of the turbine blade based on the Canny algorithm to obtain an edge image of the turbine blade and the grayscale values ​​of the pixels in the edge image, and obtaining the gradient amplitudes of all pixels in the edge image using the edge algorithm; S103, respectively correcting the grayscale values ​​of the pixels in the edge image and the gradient amplitudes of the pixels; the grayscale values ​​of the pixels in the edge image are corrected as follows: in, Indicates that it is located Grayscale correction value of the pixel at ; Indicates that it is located The original gray value of the pixel at Indicates the maximum value of the pixel grayscale value in the original grayscale image of the turbine blade; Indicates the minimum grayscale value of the pixel in the original grayscale image of the turbine blade; the gradient amplitude of the pixel in the edge image is corrected, specifically: in, Indicates that it is located The gradient amplitude correction value of the pixel point at ; Indicates that among the remaining pixels in the original edge image that have been determined as edge pixels, The original gradient amplitude of the pixel at ; Indicates the maximum value of the original gradient amplitude among the remaining pixels except those that have been determined as edge pixels in the original edge image; Indicates the minimum value of the original gradient amplitude among the remaining pixels in the original edge image except those that have been determined as edge pixels.

[0026] S104: Treat the area enclosed by closed edge pixels in the edge image as several independent large areas. For any large area, divide the area into several small areas based on the gradient amplitude correction values ​​and pixel grayscale correction values ​​of the pixels in the area other than the edge pixels of the large area. Repeat this step to divide all large areas. S104.1: Obtain judgment coefficients for the pixels in the divided large areas other than the edge pixels of the large area based on the gradient amplitude correction values ​​and pixel grayscale correction values ​​of the pixels in the area other than the edge pixels of the large area. in, Indicates that the pixels in the divided large area are located outside the edge of the large area. Judgment coefficient of position pixel point; Indicates that the pixels in the divided large area are located outside the edge of the large area. The gradient amplitude correction value of the position pixel point; Indicates that the pixels in the divided large area are excluding the edge pixels of the large area and the pixels located in The gradient amplitude correction value of the pixel points other than the position pixel point; Indicates that the pixels in the divided large area are excluding the edge pixels of the large area and the pixels located in The number of pixels outside the position pixel; Indicates that the pixels in the divided large area are located outside the edge of the large area. Grayscale correction value of the pixel at the position; Indicates that the pixels in the divided large area are excluding the edge pixels of the large area and the pixels located in Grayscale correction value of pixels other than the position pixel.

[0027] S104.2, consider the pixel with the largest judgment coefficient as the growth center pixel, and then construct a growth window. Based on the pixels between the growth windows, obtain the defect coefficient of the growth center pixel in the divided large area. Use the pixel with the largest judgment coefficient as the growth center pixel, establish an n*n window with the center of the window as the growth center pixel, and grow it according to the following rules: in, Indicates the defect coefficient of the growth center pixel in the divided large area; Represents the average value of the gradient amplitude correction value of all pixels in the window; Indicates the variance of the gradient amplitude correction values ​​of all pixels in the window; Grayscale correction value of growth center pixel in the divided large area; Indicates the average grayscale correction value of the remaining pixels in the window except the center pixel of the window.

[0028] S104.3, select any growth window as the central growth window, and the central growth window grows toward the adjacent growth window. Based on the ratio of the defect coefficients between the central growth window and the adjacent growth window, determine whether the adjacent growth window and the central growth window are in the same area; if so, select the adjacent growth window as the center, and continue to grow with other growth windows until the growth window is filled and cannot grow anymore; if not, stop growing; when the ratio of the defect coefficients between the central growth window and the adjacent growth window is greater than the set threshold, the adjacent growth window and the central growth window are considered to be in the same area; wherein, when the defect coefficients between the central growth window and the adjacent growth window are compared, the one with the larger defect coefficient is used as the denominator.

[0029] S104.4, repeat the above steps until all pixels in the large area are divided into small areas, and the small area division operation is completed.

[0030] S105, based on the pixel information in the divided small areas, obtaining a scoring coefficient for each small area, and then determining whether the current small area is a defect area; in, Indicates the The scoring coefficient of a small area; Indicates the The average value of the defect coefficient of each pixel in a small area; Indicates the minimum value of the mean defect coefficient of all small areas; Indicates the The number of pixels in a small area; determine whether the current small area is a defect area, specifically: based on the scoring coefficient of each small area obtained , get the average score coefficient of the whole picture ; If Rating coefficient of a small area Greater than the average rating coefficient of the entire map , then the corresponding small area is marked as a defect area.

[0031] S106 , if there is overlap between adjacent defective regions, the real defective regions are screened based on the overlapping areas between the adjacent defective regions.

[0032] Obtaining the ratio of the overlapping area between adjacent defect regions to the total area of ​​the corresponding two adjacent regions;

[0033] Determine whether the obtained ratio is greater than the preset threshold. If it exceeds, it means that there is a serious overlap between the two areas. In this case, the scoring coefficient is retained. The high defect areas are selected and the rest of the areas are eliminated; the real defect areas are obtained by screening.

[0034] See also Figure 2 The present invention discloses a surface defect detection system for a turbine assembly of an air cycle machine, comprising: a preprocessing module, which collects an image of a turbine blade to be detected and preprocesses the collected image; an acquisition module, which extracts edge information of the turbine blade based on a Canny algorithm, obtains an edge image of the turbine blade and the grayscale value of pixels in the edge image, and obtains the gradient amplitude of all pixels in the edge image through an edge algorithm; a correction module, which corrects the grayscale value of pixels and the gradient amplitude of pixels in the edge image respectively; a segmentation module, which divides closed edge pixels in the edge image into The area surrounded by the points is regarded as several independent large areas. For any large area, the area is divided based on the gradient amplitude correction value and the pixel grayscale correction value of the pixel points in the area except the edge pixels of the large area to obtain several small areas; and this step is repeated to divide all the large areas; a judgment module is used, which obtains the scoring coefficient of each small area based on the pixel point information in the divided small areas, and then determines whether the current small area is a defect area; a screening module is used, which screens out the real defect area based on the overlapping area between the adjacent defect areas if there is overlap between the adjacent defect areas.

[0035] Embodiment: The present invention discloses a method for detecting surface defects of a turbine assembly of an air cycle machine, comprising step 1: photographing the turbine blade to be inspected, and gray-scaling and denoising the photographed image.

[0036] During long-term use, the turbine blades of an air cycle machine may develop surface defects such as cracks, wear, and corrosion due to high temperature, high pressure, and complex airflow. At the same time, due to a period of use, dust will accumulate on the surface of the turbine blades, which will affect the detection of surface defects of the turbine blades. Therefore, the turbine blades to be inspected need to be cleaned to a certain extent, and then a clean background is selected as the shooting background of the turbine blades. A high-definition camera is used to shoot the turbine blades as close as possible while ensuring that the complete turbine blades can be photographed. The captured turbine blade image is grayscaled and denoised, and the turbofan is segmented from the image using image segmentation technology to obtain a pre-processed grayscale image of the turbine blades.

[0037] Step 2: Perform feature processing on the image data, divide the image into regions based on the processed data, and finally perform defect detection on each region to filter out the defective areas.

[0038] In step 2.1, the Canny algorithm is used to obtain the edge image of the turbine blade and the gradient amplitude of each pixel.

[0039] The cracks, wear, corrosion and other surface defects of turbine blades caused by long-term use all have certain edge features. The edge information of turbine blades is extracted using the Canny algorithm to obtain the edge image of the turbine blades, such as Figure 3 shown.

[0040] pass Figure 3 It can be seen that the edge image of the turbine blade can only obtain the general structure of the turbine blade and extremely obvious surface damage and cracks, while the small cracks, corrosion and wear on the surface cannot achieve the desired results through edge detection. Therefore, the gradient values ​​of all pixel points are obtained through the edge algorithm.

[0041] In step 2.2, the image data is processed according to the image features of the turbine blade, and the processed data is used to divide the area of ​​the blade, and finally the defect detection is performed on the divided area.

[0042] Because the overall structure and grayscale value of the turbine blade are relatively regular, as long as there is no area with particularly serious defects, the grayscale value and gradient value of the turbine blade surface defect are relatively close to the normal area. As a result, when using various data of the original image, the data of the defect area and the data of the normal area will be very close, which increases the difficulty of machine identification. Therefore, it is necessary to amplify the grayscale value difference in the turbine blade image and the gradient amplitude difference of each pixel point according to the image characteristics of the turbine blade, thereby amplifying the characteristics of all pixels. The pixel point in the lower right corner of the image is used as the coordinate origin, and the horizontal direction is to the right. The positive direction of the axis, vertical direction upward Positive direction of the axis.

[0043] Correct the grayscale value of the pixel in the edge image, specifically: in, Indicates that it is located Grayscale correction value of the pixel at Indicates that it is located The original gray value of the pixel at Indicates the maximum value of the pixel grayscale value in the original grayscale image of the turbine blade; This represents the minimum grayscale value of a pixel in the original grayscale image of the turbine blade. To amplify the gaps between pixels in the original grayscale image, the grayscale interval in the original grayscale image is directly defined as the difference between the maximum and minimum grayscale values, and the values ​​0 to 255 are redistributed within this interval. Therefore, the difference between the original grayscale image and the original grayscale image's minimum grayscale value is divided by the difference between the original grayscale image's maximum and minimum grayscale values ​​to determine the pixel's contribution to the original grayscale image. Multiplying this by 255 yields the corrected grayscale value for that pixel.

[0044] Correct the gradient amplitude of the pixel points in the edge image, specifically: in, Indicates that it is located The gradient amplitude correction value of the pixel point at ; Indicates that among the remaining pixels in the original edge image that have been determined as edge pixels, The original gradient amplitude of the pixel at ; Indicates the maximum value of the original gradient amplitude among the remaining pixels except those that have been determined as edge pixels in the original edge image; Indicates the minimum value of the original gradient amplitude among the remaining pixels in the original edge image except those that have been determined as edge pixels.

[0045] In the original edge image, due to the structure of the turbine blades, the strong edges in the original edge image are all caused by the structure, and the gradient amplitude of the structure is particularly large. If these edge pixels are also involved in the gradient correction of other pixels, it will greatly affect the accuracy of the correction, so that the gradient amplitude after correction does not show its characteristics better. Therefore, the edge pixels in the original edge image are removed during the correction process. And the original gradient value range is 0 to Therefore, in the process of gradient amplitude correction, 361 is redistributed according to its original gradient amplitude ratio. At this point, the grayscale correction value and gradient amplitude correction value after correction based on the characteristics of the turbine blade are obtained.

[0046] It can be clearly seen from the image of the turbine blade that due to its structural problems, the regionalization of the turbine blade is very obvious. There are strong regional characteristics between different blades, as well as between the blades and the central areas connecting the blades. They are independent individuals. Therefore, the turbine blades can be regionalized first during defect detection.

[0047] First, based on the distribution characteristics of edge pixels in the original edge image of the turbine blade, the areas enclosed by closed edge pixels in the edge image are considered as independent large areas. If defect detection is performed directly on these large areas, the detection process will be relatively rough, and the detection results will easily overlook some relatively subtle defects. Therefore, these large areas need to be further divided. According to the defect characteristics of the turbine blade, whether it is a tiny crack or an area that is corroded or worn, its surface is uneven compared to the surrounding blades without problems, and these defective areas are exposed to their internal structure because the surface structure is destroyed, resulting in differences in grayscale values ​​from normal areas. Therefore, each large area can be further divided according to these characteristics.

[0048] When dividing these large areas into small areas, we first need to find the pixels that are more likely to be defective areas, and then divide one of the large areas based on these pixels. Specifically, based on the gradient amplitude correction value and pixel grayscale correction value of the pixels in the area except the edge pixels of the large area, we obtain the judgment coefficients of the pixels in the divided large area except the edge pixels of the large area. Specifically, in, Indicates that the pixels in the divided large area are located outside the edge of the large area. Judgment coefficient of position pixel point; Indicates that the pixels in the divided large area are located outside the edge of the large area. The gradient amplitude correction value of the position pixel; Indicates that the pixels in the divided large area are excluding the edge pixels of the large area and the pixels located in The gradient amplitude correction value of the pixel points other than the position pixel point; Indicates that the pixels in the divided large area are excluding the edge pixels of the large area and the pixels located in The number of pixels outside the position pixel; Indicates that the pixels in the divided large area are located outside the edge of the large area. Grayscale correction value of the pixel at the position; Indicates that the pixels in the divided large area are excluding the edge pixels of the large area and the pixels located in Grayscale correction value of pixels other than the position pixel.

[0049] When dividing these large areas into small areas, we must first find the pixels that are more likely to be defective areas, and use these pixels as the basis for area division. The most obvious feature of the pixels that may be defective areas is that the grayscale value and gradient amplitude of the pixel are somewhat different from those of the normal area. Therefore, the grayscale values ​​and gradient amplitudes of all pixels in the divided area and the remaining pixels are summed and averaged to obtain the judgment coefficient of each pixel in the large area. The larger the judgment coefficient, the more likely it is that the pixel is in the defective area.

[0050] After obtaining the judgment coefficients of all pixels in the divided large area except for the edge pixels, the pixel with the largest judgment coefficient is used as the growth center pixel. A 3*3 window is established with the center of the window as the growth center pixel, and growth is performed according to the following rules: in, Indicates the defect coefficient of the growth center pixel in the divided large area; Represents the average value of the gradient amplitude correction value of all pixels in the window; Indicates the variance of the gradient amplitude correction values ​​of all pixels in the window; Grayscale correction value of growth center pixel in the divided large area; Represents the average grayscale correction value of the remaining pixels in the window except the pixel at the center of the window; the difference between the grayscale value and the gradient amplitude between the defective area and the normal area of ​​the turbine blade. Therefore, the pixels that may be defective areas are screened based on these two features. Because the surface of the defective area is rougher than that of the normal area, if the average gradient amplitude correction value of the pixels in the window is relatively small, but the variance is relatively large, it means that the gradient amplitude correction value difference between the pixels in this window is extremely large, then this window is likely to contain an area with surface defects, and the defect coefficient of the growth center pixel at the center of the window will be larger. If the grayscale correction value difference between the growth center pixel at the center of the window and the rest of the pixels in the window is too large, it means that the difference between the center pixel and the surrounding pixels is too large, which will increase the defect coefficient of the center pixel, thereby increasing the possibility that the window area may be a defective area.

[0051] After establishing a window with the pixel with the largest judgment coefficient as the growth center pixel, growth is performed in eight directions. The growth diagram is shown in the following figure. Figure 4 shown.

[0052] For example, if the initial growth window is A, then growth is performed in the other eight directions. When growing toward window B, if the ratio of the defect coefficients of window A to window B (the larger defect coefficient is used as the denominator) is greater than or equal to 0.8, window B is marked as the same region as window A. If it is less than 0.8, they are not marked as the same region and growth stops. If window B has already been marked as the same region, growth continues outward from window B using the same method until growth cannot be continued.

[0053] When a small area is completely unable to grow, except for this small area, continue to use the judgment coefficient The largest pixel is used as the growth center pixel, and the above method is used to continue growing until all pixels in the large area are divided into small areas.

[0054] According to the defect coefficient After the small areas are divided, all the small areas now contain pixels with very similar features. Therefore, if there are defects in some areas, the defective pixels will be grouped in the same small area, while the normal turbine blade areas will be concentrated in one area. Therefore, a final judgment is required between the defective areas and the normal areas: in, Indicates the The scoring coefficient of a small area; Indicates the The average value of the defect coefficient of each pixel in a small area; Indicates the minimum value of the mean defect coefficient in all small areas (referring to all small areas contained in the image, not just all small areas in a large area); Indicates the The number of pixels in a small area; the defective area of ​​the turbine blade to be inspected is a very small area relative to the normal area of ​​the entire turbine blade. Therefore, the number of pixels in the small area that may be the defective area is very small relative to the normal area. At the same time, the defect coefficient of the pixels in the defective area is relatively large relative to the normal area. Therefore, if the first If an area is a defective area, the average defect coefficient of the area is relatively large, and then subtracting the minimum value of the average defect coefficient will produce a very large difference. In addition, the number of pixels in the defective area is very small compared to the normal area, so the scoring coefficient of the defective area will be large, while the scoring coefficient of the normal area will be relatively small. The gap between the two will be very obvious. Therefore, Greater than the average rating coefficient of the entire map The area is marked as a defective area. So far, the defective areas in all small areas are distinguished from the normal areas.

[0055] In step 2.3, the detected defect areas are subjected to overlapping maximum suppression to screen out the real defect areas.

[0056] In the step of dividing small areas, some small areas may be in an overlapping state, and some of these overlapping small areas may also overlap with defective areas, so these overlapping areas need to be screened.

[0057] Obtain the overlapping area between all marked defect areas, and compare the overlapping area with the total area of ​​the two areas. If the ratio is greater than 0.5, it means that there is a serious overlap between the two areas, and the scoring coefficient is retained. The highest area is selected, and the rest of the areas are eliminated; at this time, the defect area left is the most complete and accurate defect area.

[0058] An embodiment of the present invention provides a terminal device. The terminal device of this embodiment includes: a processor, a memory, and a computer program stored in the memory and executable on the processor. When the processor executes the computer program, the steps of each of the aforementioned method embodiments are implemented. Alternatively, when the processor executes the computer program, the functions of each module / unit in each of the aforementioned device embodiments are implemented.

[0059] The computer program may be divided into one or more modules / units, which are stored in the memory and executed by the processor to accomplish the present invention.

[0060] The terminal device may be a computing device such as a desktop computer, a notebook computer, a PDA, a cloud server, etc. The terminal device may include, but is not limited to, a processor and a memory.

[0061] The processor can be a central processing unit (CPU), or other general-purpose processors, digital signal processors (DSP), application-specific integrated circuits (ASIC), field-programmable gate arrays (FPGA) or other programmable logic devices, discrete gate or transistor logic devices, discrete hardware components, etc.

[0062] The memory may be used to store the computer programs and / or modules, and the processor implements various functions of the terminal device by running or executing the computer programs and / or modules stored in the memory and calling the data stored in the memory.

[0063] If the module / unit integrated into the terminal device is implemented as a software functional unit and sold or used as an independent product, it can be stored in a computer-readable storage medium. Based on this understanding, the present invention can implement all or part of the process steps in the above-mentioned method embodiments by using a computer program to instruct the relevant hardware. The computer program can be stored in a computer-readable storage medium. When executed by a processor, the computer program can implement the steps of each of the above-mentioned method embodiments. The computer program includes computer program code, which can be in source code form, object code form, executable file, or some intermediate form. The computer-readable medium can include: any entity or device capable of carrying the computer program code, recording medium, USB flash drive, mobile hard drive, magnetic disk, optical disk, computer memory, read-only memory (ROM), random access memory (RAM), electric carrier signal, telecommunication signal, and software distribution medium. It should be noted that the content of the computer-readable medium can be appropriately increased or decreased based on the requirements of legislation and patent practice in a jurisdiction. For example, in some jurisdictions, according to legislation and patent practice, computer-readable media does not include electric carrier signals and telecommunication signals.

[0064] The above are merely preferred embodiments of the present invention and are not intended to limit the present invention. Those skilled in the art will readily appreciate that various modifications and variations of the present invention are possible. Any modifications, equivalent substitutions, or improvements made within the spirit and principles of the present invention shall be included within the scope of protection of the present invention.

Claims

1. A method for detecting surface defects of a turbine assembly of an air cycle machine, characterized in that: include: Collecting images of turbine blades to be inspected and preprocessing the collected images; The edge information of the turbine blade is extracted based on the Canny algorithm to obtain the edge image of the turbine blade and the grayscale value of the pixel points in the edge image. The gradient amplitude of all pixel points in the edge image is obtained through the edge algorithm. Correct the grayscale value and gradient amplitude of the pixel points in the edge image respectively; The area enclosed by closed edge pixels in the edge image is considered as several independent large areas. For any large area, the area is divided based on the gradient amplitude correction value and pixel grayscale correction value of the pixels in the area except the edge pixels of the large area to obtain several small areas. This step is repeated to divide all the large areas. Obtaining, based on the gradient amplitude correction values ​​and pixel grayscale correction values ​​of the pixels in the region other than the edge pixels of the large region, the determination coefficients of the pixels in the divided large region other than the edge pixels of the large region; in, Indicates that the pixels in the divided large area are located except for the edge pixels of the large area. Judgment coefficient of position pixel point; Indicates that the pixels in the divided large area are located except for the edge pixels of the large area. The gradient amplitude correction value of the position pixel; Indicates that the pixels in the divided large area are excluding the edge pixels of the large area and the pixels located in The gradient amplitude correction value of the pixel points other than the position pixel point; Indicates that the pixels in the divided large area are excluding the edge pixels of the large area and the pixels located in The number of pixels outside the position pixel; Indicates that the pixels in the divided large area are located except for the edge pixels of the large area. Grayscale correction value of the pixel at the position; Indicates that the pixels in the divided large area are excluding the edge pixels of the large area and the pixels located in Grayscale correction value of pixels other than the position pixel; The pixel with the largest judgment coefficient is regarded as the growth center pixel, and then a growth window is constructed. Based on the pixels between the growth windows, the defect coefficient of the growth center pixel in the divided large area is obtained. Specifically, the pixel with the largest judgment coefficient is regarded as the growth center pixel, and an n*n window is established with the center of the window as the growth center pixel. The growth is performed according to the following rules: in, Indicates the defect coefficient of the growth center pixel in the divided large area; Represents the average value of the gradient amplitude correction value of all pixels in the window; Indicates the variance of the gradient amplitude correction values ​​of all pixels in the window; Grayscale correction value of growth center pixel in the divided large area; Indicates the average grayscale correction value of the remaining pixels in the window except the center pixel of the window; Based on the pixel information in the divided small areas, the scoring coefficient of each small area is obtained, and then whether the current small area is a defect area is determined; If there is overlap between adjacent defective areas, the real defective areas are screened out based on the overlapping areas between the adjacent defective areas.

2. A method for detecting surface defects of a turbine assembly for an air cycle machine according to claim 1, characterized in that: The preprocessing of the collected image is specifically: grayscale denoising is performed on the collected turbine blade image, and the turbofan is segmented from the image using image segmentation technology to obtain a preprocessed turbine blade grayscale image.

3. The method for detecting surface defects of a turbine assembly for an air cycle machine according to claim 2, characterized in that: The grayscale value of the pixel in the edge image is corrected as follows: in, Indicates that it is located Grayscale correction value of the pixel at ; Indicates that it is located The original gray value of the pixel at Indicates the maximum value of the pixel grayscale value in the original grayscale image of the turbine blade; Indicates the minimum grayscale value of the pixel in the original grayscale image of the turbine blade; the gradient amplitude of the pixel in the edge image is corrected, specifically: in, Indicates that it is located The gradient amplitude correction value of the pixel point at ; Indicates that among the remaining pixels in the original edge image that have been determined as edge pixels, The original gradient amplitude of the pixel at ; Indicates the maximum value of the original gradient amplitude among the remaining pixels in the original edge image except those that have been determined as edge pixels; Indicates the minimum value of the original gradient amplitude among the remaining pixels in the original edge image except those that have been determined as edge pixels.

4. The method for detecting surface defects of a turbine assembly for an air cycle machine according to claim 3, wherein: The original grayscale image is divided into several independent large areas with the closed edge pixels in the edge image as the boundary. For any large area, the area is divided based on the gradient amplitude correction value and the grayscale correction value of the pixels in the area except the edge pixels of the large area to obtain several small areas. This step is repeated to divide all the large areas into regions. Specifically: The pixel point with the largest judgment coefficient is regarded as the growth center pixel point, and then the growth window is constructed; based on the pixel points between the growth windows, the defect coefficient of the growth center pixel point in the divided large area is obtained; any growth window is selected as the central growth window, and the central growth window grows toward the adjacent growth window, and based on the ratio of the defect coefficient between the central growth window and the adjacent growth window, it is judged whether the adjacent growth window and the central growth window are in the same area; if so, the adjacent growth window is selected as the center, and continues to grow with other growth windows until the growth window is filled and cannot grow; if not, the growth is stopped; repeat the above steps until all pixel points in the large area are divided into small areas, and the small area division operation is completed.

5. The method for detecting surface defects of a turbine component of an air cycle machine according to claim 4, characterized in that: The method judges whether the adjacent growth window and the central growth window are in the same area based on the ratio of the defect coefficient between the central growth window and the adjacent growth window. Specifically, when the ratio of the defect coefficient between the central growth window and the adjacent growth window is greater than the set threshold, the adjacent growth window and the central growth window are considered to be in the same area; wherein, when the defect coefficients between the central growth window and the adjacent growth window are compared, the one with the larger defect coefficient is used as the denominator.

6. The method for detecting surface defects of a turbine component of an air cycle machine according to claim 5, characterized in that: The scoring coefficient of each small area is obtained based on the pixel information in the divided small area, specifically: in, Indicates the The scoring coefficient of a small area; Indicates the The average value of the defect coefficient of each pixel in a small area; Indicates the minimum value of the mean defect coefficient of all small areas; Indicates the The number of pixels in a small area; the judgment of whether the current small area is a defect area is specifically: based on the obtained scoring coefficient of each small area , get the average score coefficient of the whole picture ; If Rating coefficient of a small area Greater than the average rating coefficient of the entire map , then the corresponding small area is marked as a defect area.

7. The method for detecting surface defects of a turbine component of an air cycle machine according to claim 6, characterized in that: The method of screening out the real defective areas based on the overlapping areas between adjacent defective areas is as follows: obtaining the ratio of the overlapping areas between adjacent defective areas to the total area of ​​the two corresponding adjacent areas; judging whether the obtained ratio is greater than a preset threshold; if so, it indicates that there is a serious overlap between the two areas, and the scoring coefficient is retained. The high defect areas are selected and the rest of the areas are eliminated; the real defect areas are obtained by screening.

8. A surface defect detection system for a turbine assembly of an air cycle machine, characterized in that: include: A preprocessing module, wherein the preprocessing module collects images of turbine blades to be inspected and preprocesses the collected images; An acquisition module extracts edge information of the turbine blade based on a Canny algorithm, obtains an edge image of the turbine blade and the grayscale values ​​of pixels in the edge image, and obtains the gradient amplitude of all pixels in the edge image through an edge algorithm; A correction module, wherein the correction module corrects the grayscale value and gradient amplitude of the pixel points in the edge image respectively; A division module, wherein the division module regards the area enclosed by closed edge pixels in the edge image as several independent large areas. For any large area, the area is divided based on the gradient amplitude correction value and the pixel grayscale correction value of the pixel points in the area except the edge pixels of the large area, to obtain several small areas; and this step is repeated to divide all the large areas; Obtaining, based on the gradient amplitude correction values ​​and pixel grayscale correction values ​​of the pixels in the region other than the edge pixels of the large region, the determination coefficients of the pixels in the divided large region other than the edge pixels of the large region; in, Indicates that the pixels in the divided large area are located except for the edge pixels of the large area. Judgment coefficient of position pixel point; Indicates that the pixels in the divided large area are located except for the edge pixels of the large area. The gradient amplitude correction value of the position pixel; Indicates that the pixels in the divided large area are excluding the edge pixels of the large area and the pixels located in The gradient amplitude correction value of the pixel points other than the position pixel point; Indicates that the pixels in the divided large area are excluding the edge pixels of the large area and the pixels located in The number of pixels outside the position pixel; Indicates that the pixels in the divided large area are located except for the edge pixels of the large area. Grayscale correction value of the pixel at the position; Indicates that the pixels in the divided large area are excluding the edge pixels of the large area and the pixels located in Grayscale correction value of pixels other than the position pixel; The pixel with the largest judgment coefficient is regarded as the growth center pixel, and then a growth window is constructed. Based on the pixels between the growth windows, the defect coefficient of the growth center pixel in the divided large area is obtained. Specifically, the pixel with the largest judgment coefficient is regarded as the growth center pixel, and an n*n window is established with the center of the window as the growth center pixel. The growth is performed according to the following rules: in, Indicates the defect coefficient of the growth center pixel in the divided large area; Represents the average value of the gradient amplitude correction value of all pixels in the window; Indicates the variance of the gradient amplitude correction values ​​of all pixels in the window; Grayscale correction value of growth center pixel in the divided large area; Indicates the average grayscale correction value of the remaining pixels in the window except the center pixel of the window; A judgment module, which obtains a scoring coefficient for each small area based on pixel information in the divided small areas, and then determines whether the current small area is a defective area; A screening module is provided. If there is overlap between adjacent defective areas, the screening module screens out the real defective areas based on the overlapping areas between the adjacent defective areas.

Images