Image enhancement method for quality detection of self-priming pump impeller

Abstract

The invention discloses an image enhancement method for quality detection of a self-priming pump impeller, and the method comprises the steps: S1, collecting a surface detection image of the self-priming pump impeller, and carrying out the image partitioning of the surface detection image of the self-priming pump impeller; s2, optimizing the CLAHE equalization process through the cavitation degree of pixel points in the image, and obtaining a gray value mapping result of the CLAHE process; according to the cavitation degree of the pixel point, the cavitation intensity corresponding to each gray value is evaluated; after the cavitation intensity of the gray histogram in the sub-block is obtained, the gray value cumulative distribution function of the sub-block is optimized according to the cavitation intensity corresponding to the gray value, so that the gray value mapping is optimized, and the gray value mapping result of the CLAHE process is obtained. The method relates to the technical field of self-priming pump impeller quality detection, and has the beneficial effects that the image on the surface of the self-priming pump impeller is enhanced to eliminate the influence of metal reflection and highlight the image characteristics of cavitation defects, so that the accuracy of impeller quality detection is improved.

Description

Technical Field

[0001] The invention relates to the field of self-priming pump impeller quality detection, in particular to an image enhancement method for self-priming pump impeller quality detection. Background Art

[0002] As the core component of a self-priming pump, the impeller is responsible for sucking in liquid and driving the liquid to flow. The quality of the impeller directly affects the efficiency, stability and service life of the pump. The surface of the impeller may be damaged due to long-term use, overload, liquid impact or cavitation, among which cavitation defects are particularly serious, which may cause tiny holes or corrosion on the surface of the impeller, thus affecting the performance of the pump. Therefore, accurate detection of impeller surface defects (especially cavitation defects) is crucial to ensure the safe operation of the pump. Summary of the invention

[0003] The purpose of the present invention is to solve the above problems and to design an image enhancement method for quality detection of a self-priming pump impeller.

[0004] The technical solution of the present invention to achieve the above-mentioned purpose is an image enhancement method for self-priming pump impeller quality detection, comprising:

[0005] Step S1: collecting a surface detection image of a self-priming pump impeller, and dividing the surface detection image of the self-priming pump impeller into image blocks;

[0006] Step S2: Optimizing the CLAHE equalization process by the cavitation degree of the pixel points in the image to obtain the gray value mapping result of the CLAHE process;

[0007] The cavitation degree of the pixel point is obtained through the local texture information of the pixel point;

[0008] The cavitation intensity corresponding to each gray value is evaluated according to the cavitation degree of the pixel point;

[0009] After obtaining the cavitation intensity of the grayscale histogram in the sub-block, the grayscale value cumulative distribution function of the sub-block is optimized according to the cavitation intensity corresponding to the grayscale value, thereby optimizing the grayscale value mapping and obtaining the grayscale value mapping result of the CLAHE process;

[0010] Step S3: The grayscale value of the self-priming pump impeller surface detection image is changed through the grayscale value mapping result of the CLAHE process to complete the image enhancement.

[0011] Furthermore, in step S1: a high-definition industrial camera collects the surface image of the self-priming pump impeller, converts the collected image into a grayscale image and crops the surface image of the self-priming pump impeller to the same size. After collecting the surface detection image of the self-priming pump impeller, the surface detection image of the self-priming pump impeller needs to be segmented.

[0012] Further, in step S2: by evaluating the cavitation degree of the pixel points in the self-priming pump impeller surface detection image (grayscale image), the cavitation degree of the pixel points in the image is evaluated according to the local texture information of the pixel points in the image. After the cavitation degree of the pixel points in the image is obtained, the grayscale histogram evaluation process of each image sub-block can be optimized according to the cavitation degree of the pixel points;

[0013] In the statistical process of the grayscale histogram, the cavitation intensity corresponding to each grayscale value is evaluated according to the cavitation degree of the pixel. This cavitation intensity is the degree to which there are cavitation defective pixels in the pixel corresponding to the grayscale value. The higher the cavitation intensity, the more it is necessary to evaluate the grayscale mapping result according to the cavitation intensity of each grayscale value in the evaluation of the grayscale value mapping, so as to ensure that the local pixel value difference relationship of the defective pixel in the image will not be eliminated in the process of histogram equalization by limiting the contrast. After obtaining the cavitation intensity of the grayscale histogram in the sub-block, the grayscale value cumulative distribution function of the sub-block is optimized according to the cavitation intensity corresponding to the grayscale value, so as to optimize the grayscale value mapping and obtain the grayscale value mapping result of the CLAHE process.

[0014] Furthermore, the cavitation degree of the pixel point is obtained through the local texture information of the pixel point; after obtaining the image sub-block of the self-priming pump impeller surface detection image, the cavitation degree of each pixel point in the image is first evaluated. Because when cavitation defects appear on the surface of the self-priming pump impeller, they will appear as irregular tiny pits in the image, so the cavitation degree of the central pixel point is evaluated by setting the 24 neighborhoods of the pixel point in the image (that is, the local 5*5 window of each pixel point). Cavitation degree of each pixel:

[0015]

[0016] Explanation of formula symbols:

[0017] :Indicates the first The local window of the pixel The gray value of a pixel.

[0018] :Indicates the first The local window of the pixel Pixels in The grayscale value of the adjacent pixel points on the right side of the axis.

[0019] :Indicates the first The local window of the pixel Pixels in Grayscale values ​​of adjacent pixels on the axis.

[0020] :Indicates the first The number of pixels in a local window of pixels.

[0021] :Indicates the first The local window of the pixel The gradient direction of each pixel.

[0022] :Indicates the first The local window of the pixel The frequency with which the gradient direction of a pixel appears in the gradient directions of all pixels in the window.

[0023] : represents the linear normalization function.

[0024] Furthermore, the cavitation intensity corresponding to each gray value in the sub-block is obtained by the cavitation degree of all pixels in the sub-block;

[0025] After obtaining the cavitation degree of each pixel in the image, the overall cavitation intensity of all pixels corresponding to each gray value can be evaluated according to the cavitation degree of the pixel in the image sub-block. In the image sub-block, when the overall cavitation degree of the pixel corresponding to a gray value is higher, the cavitation intensity of the gray value is also higher. Therefore, in the subsequent gray value mapping process, the gray value mapping process can be optimized by the cavitation intensity of the gray value, and the difference relationship between gray values ​​with higher cavitation intensity is retained. Gray value of cavitation intensity :

[0026]

[0027] Explanation of formula symbols:

[0028] : represents the first The number of pixels corresponding to a gray value.

[0029] Further, the gray value cumulative distribution function is optimized by the cavitation intensity of the pixel value in the sub-block, and an optimized gray value mapping result is obtained according to the optimized cumulative distribution function;

[0030] After obtaining the cavitation intensity of each gray value in the image sub-block, the mapping optimization factor corresponding to the gray value can be obtained through the cavitation intensity relationship between the gray values, so that the original gray value is optimized in the gray value mapping evaluation process through the cumulative distribution function through the gray value mapping optimization factor. In the mapping evaluation process, the difference between gray values ​​with higher cavitation intensity is retained, that is, it is avoided to map several gray values ​​with higher cavitation intensity to gray values ​​with smaller differences. Gray value mapping optimization factor :

[0031]

[0032] In the above formula, the cavitation intensity of the gray value is of Normalization is used to evaluate the mapping optimization factor. In the process of grayscale value mapping, because the relative relationship between pixels needs to be retained, the mapping optimization factor needs to be evaluated through the difference information of all grayscale values, so as to ensure that the difference between grayscale values ​​with large differences is retained in the subsequent grayscale value mapping process. And because the pixel value difference between grayscale values ​​with higher cavitation intensity needs to be retained in the process of regulating the cumulative distribution function by the mapping optimization factor, the mapping optimization factor between grayscale values ​​with higher cavitation intensity and grayscale values ​​with lower cavitation intensity needs to be enhanced to highlight the difference between grayscale values ​​with higher cavitation intensity and the surrounding pixels. The above formula is obtained by Normalization is used to enhance the difference between the mapping optimization factors between the grayscale values ​​with higher cavitation intensity and the grayscale values ​​with lower cavitation intensity, thereby ensuring that the optimized grayscale value mapping can accurately reflect the local grayscale value difference information of the cavitation defect pixels in the image.

[0033] After obtaining the mapping optimization factor of the gray value, the mapping optimization factor can be used to evaluate in the mapping evaluation process to obtain the first The mapping target gray value of gray value :

[0034]

[0035] Formula explanation:

[0036] :Indicates the first The mapping optimization factor of gray values;

[0037] : No. The cumulative distribution function value corresponding to the gray value;

[0038] : Indicates the minimum value of the number of pixels corresponding to the grayscale value in the image sub-block, that is, the minimum value of the cumulative distribution function.

[0039] : Respectively represent the length and width of the image sub-block, and the unit is the number of pixels of the length and width respectively. In the present invention, ;

[0040] : Indicates the number of gray values ​​in the image sub-block;

[0041] In the above formula, through The gray value mapping optimization factor optimizes the gray value mapping evaluation process of CLAHE. The higher the mapping optimization factor, the better the gray value mapping evaluation process. The closer the mapping target gray value of a gray value is to other gray values ​​with higher cavitation intensity;

[0042] In the above formula, through As the The weight of the cumulative distribution function of the gray value, so when The higher the value of The mapping effect corresponding to each gray value changes more significantly, thereby preserving the differences between pixels during the mapping process.

[0043] Furthermore, after obtaining the gray value mapping result of the CLAHE process, the pixel value of each image sub-block can be adjusted according to the gray value mapping result, thereby completing the image sub-block equalization process. After obtaining the mapping of the pixel value of each sub-block of the self-priming pump impeller surface detection image, all image sub-blocks can be integrated by bilinear interpolation to obtain the final image enhancement processing result.

[0044] An image enhancement method for self-priming pump impeller quality inspection is made using the technical solution of the present invention. The image of the self-priming pump impeller surface is enhanced to eliminate the influence of metal reflection and highlight the image characteristics of cavitation defects, thereby improving the accuracy of impeller quality inspection. The method mainly optimizes the traditional CLAHE image enhancement technology through the following steps:

[0045] Cavitation degree assessment: The degree of cavitation defects at each pixel in the image is assessed by calculating the local gradient and texture features of each pixel. Cavitation areas usually show subtle changes in texture, so this method accurately assesses the degree of cavitation at each pixel through local texture analysis and gradient information of the image.

[0046] Optimizing the Cumulative Distribution Function (CDF): The traditional CLAHE method uses the grayscale histogram of the image for equalization. In this technical solution, the cumulative distribution function (CDF) is adjusted by weighting the evaluation results of the cavitation degree, so that the cavitation characteristics can be better retained during the grayscale mapping process of the cavitation defect area, avoiding the cavitation pixels from being over-smoothed or eliminated.

[0047] Enhance image contrast and highlight cavitation defects: By mapping the grayscale value of the image through the optimized CDF, the grayscale value change of the cavitation defect area in the enhanced image is more significant, thereby improving the visibility of the defect. At the same time, by controlling the mapping intensity of the cavitation defect area, the influence of metal surface reflection can be avoided to a certain extent, ensuring the effect of image enhancement. BRIEF DESCRIPTION OF THE DRAWINGS

[0048] Figure 1 It is a schematic diagram of an image enhancement method for quality detection of a self-priming pump impeller according to the present invention. DETAILED DESCRIPTION

[0049] The present invention will be described in detail below in conjunction with the accompanying drawings. Figure 1 As shown, an image enhancement method for quality inspection of a self-priming pump impeller, step S1: collecting a surface inspection image of the self-priming pump impeller, and dividing the surface inspection image of the self-priming pump impeller into image blocks.

[0050] The surface image of the self-priming pump impeller is collected by a high-definition industrial camera, the collected image is converted into a grayscale image and the surface image of the self-priming pump impeller is cropped to the same size. In this application, the image size is cropped to 1024*1024, and the fixed size of the image can be adjusted according to the actual scene without any requirement. After the surface detection image of the self-priming pump impeller is collected, the self-priming pump impeller surface detection image needs to be segmented, so that the self-priming pump impeller surface detection image is enhanced through the CLAHE histogram equalization process. For the image segmentation process, the image sub-block size is set to 32*32 in the present invention, and the size of the image sub-block can be adjusted according to the actual scene, and there is no requirement.

[0051] At this point, the image acquisition is completed and the acquired image is divided into blocks.

[0052] Step S2: Optimize the CLAHE equalization process by the cavitation degree of the pixel points in the image to obtain the gray value mapping result of the CLAHE process.

[0053] Big step logic:

[0054] After obtaining the sub-block division results of the self-priming pump impeller surface detection image, the grayscale histogram of each acquired image sub-block can be obtained, and the grayscale value mapping evaluation can be performed according to the cumulative distribution function corresponding to the grayscale histogram, and the grayscale value mapping result can be obtained to complete the grayscale value mapping result based on the histogram equalization based on the limited contrast, so as to perform grayscale value mapping on each sub-block and integrate all the sub-blocks into a complete image through bilinear interpolation, and the acquired complete image is used as the enhanced image obtained by the CLAHE algorithm.

[0055] In the above-mentioned process of enhancing the image by the CLAHE algorithm, the influence of the metal surface reflection appearing in the self-priming pump impeller surface detection image can be eliminated. However, in the above process, all the pixels corresponding to each gray value in the original image are transformed into other gray values ​​through gray value mapping during the image enhancement process. For the pixels that may exist in the image and represent the surface defects of the self-priming pump impeller, the difference information between some pixels will be eliminated during the gray value mapping process.

[0056] In response to the above situation, the present invention evaluates the cavitation degree of pixel points in the self-priming pump impeller surface detection image (grayscale image), and evaluates the cavitation degree of pixel points in the image according to the local texture information of the pixel points in the image. After obtaining the cavitation degree of the pixel points in the image, the grayscale histogram evaluation process of each image sub-block can be optimized according to the cavitation degree of the pixel points. Specifically, in the statistical process of the grayscale histogram, the cavitation intensity corresponding to each grayscale value is evaluated according to the cavitation degree of the pixel points.

[0057] This cavitation intensity is the degree to which there are cavitation defective pixels in the pixels corresponding to the gray value. The higher the cavitation intensity, the more it is necessary to evaluate the gray value mapping result according to the cavitation intensity of each gray value in the evaluation of gray value mapping, so as to ensure that the local pixel value difference relationship of the defective pixels in the image will not be eliminated in the process of histogram equalization by limiting the contrast. After obtaining the cavitation intensity of the gray histogram in the sub-block, the gray value cumulative distribution function of the sub-block is optimized according to the cavitation intensity corresponding to the gray value, thereby optimizing the gray value mapping and obtaining the gray value mapping result of the CLAHE process.

[0058] The cavitation degree of the pixel is obtained through the local texture information of the pixel.

[0059] Detailed logic:

[0060] After obtaining the image sub-blocks of the self-priming pump impeller surface detection image, the cavitation degree of each pixel in the image is first evaluated. Because when cavitation defects occur on the surface of the self-priming pump impeller, they will appear as irregular tiny pits in the image. Therefore, the 24 neighborhoods of the pixel points (i.e. the local area of ​​each pixel point) are set in the image. The cavitation degree of the central pixel is evaluated by using the window. Cavitation degree of each pixel:

[0061]

[0062] Explanation of formula symbols:

[0063] :Indicates the first The local window of the pixel The gray value of a pixel.

[0064] :Indicates the first The local window of the pixel Pixels in The grayscale value of the adjacent pixel points on the right side of the axis.

[0065] :Indicates the first The local window of the pixel Pixels in Grayscale values ​​of adjacent pixels on the axis.

[0066] :Indicates the first The number of pixels in a local window of pixels.

[0067] :Indicates the first The local window of the pixel The gradient direction of each pixel.

[0068] :Indicates the first The local window of the pixel The frequency with which the gradient direction of a pixel appears in the gradient directions of all pixels in the window.

[0069] : represents the linear normalization function.

[0070] Formula logic:

[0071] In the above formula, the cavitation degree of the pixel point is evaluated through the local window of the pixel point in the image. For the pixel point in the image, when the overall gradient intensity in its local window (the 24 neighborhoods mentioned above) is higher, the window is more likely to contain detail information in the image, and this part of the detail information may be the cavitation defect on the surface of the self-priming pump impeller. Because the gradient intensity cannot fully characterize the possibility that the pixel point is a cavitation defect, the cavitation degree of the pixel point is further characterized by the gradient direction information of the pixel point in the local window, that is, for the pixel point in the image, if the gradient intensity in its local window is higher and the gradient direction in the local window is more irregular, the cavitation degree of the pixel point is higher. The principle here is that the cavitation defect on the surface of the self-priming pump impeller will appear as irregular small pits, then the gradient intensity of the edge pixel point of the pit will appear at a higher level, and because of the irregularity of the pit, the gradient direction of the pixel point in the local window of the cavitation defect pixel point will appear irregular, so the cavitation degree of the pixel point is characterized by the numerical information of the entropy value of the gradient direction. In the above formula Part of it is to evaluate the gradient strength of the pixel points in the local window. The first part is to evaluate the irregularity of the gradient direction of the pixel in the local window, and the normalized result of the multiplication of the two is used as the cavitation degree of the pixel in the image.

[0072] The cavitation intensity corresponding to each gray value in the sub-block is obtained through the cavitation degree of all pixels in the sub-block.

[0073] Detailed logic:

[0074] After obtaining the cavitation degree of each pixel in the image, the overall cavitation intensity of all pixels corresponding to each gray value can be evaluated according to the cavitation degree of the pixel in the image sub-block. In the image sub-block, when the overall cavitation degree of the pixel corresponding to a gray value is higher, the cavitation intensity of the gray value is also higher. Therefore, in the subsequent gray value mapping process, the gray value mapping process can be optimized by the cavitation intensity of the gray value, and the difference relationship between gray values ​​with higher cavitation intensity is retained. Gray value of cavitation intensity :

[0075]

[0076] Explanation of formula symbols:

[0077] : represents the first The number of pixels corresponding to a gray value.

[0078] Formula logic:

[0079] In the above formula, the average cavitation degree of all pixels corresponding to each gray value is taken as the cavitation intensity corresponding to the gray value. For a gray value, the higher the average cavitation degree of the pixels in the pixel of the gray value, the higher the cavitation intensity corresponding to the gray value.

[0080] The gray value cumulative distribution function is optimized by the cavitation intensity of the pixel value in the sub-block, and the optimized gray value mapping result is obtained according to the optimized cumulative distribution function.

[0081] Detailed logic:

[0082] After obtaining the cavitation intensity of each gray value in the image sub-block, the mapping optimization factor corresponding to the gray value can be obtained through the cavitation intensity relationship between the gray values, so that the original gray value is optimized in the gray value mapping evaluation process through the cumulative distribution function through the gray value mapping optimization factor. In the mapping evaluation process, the difference between gray values ​​with higher cavitation intensity is retained, that is, it is avoided to map several gray values ​​with higher cavitation intensity to gray values ​​with smaller differences. Gray value mapping optimization factor :

[0083]

[0084] In the above formula, the cavitation intensity of the gray value is of Normalization is used to evaluate the mapping optimization factor. In the process of grayscale value mapping, because the relative relationship between pixels needs to be retained, the mapping optimization factor needs to be evaluated through the difference information of all grayscale values, so as to ensure that the difference between grayscale values ​​with large differences is retained in the subsequent grayscale value mapping process. And because the pixel value difference between grayscale values ​​with higher cavitation intensity needs to be retained in the process of regulating the cumulative distribution function by the mapping optimization factor, the mapping optimization factor between grayscale values ​​with higher cavitation intensity and grayscale values ​​with lower cavitation intensity needs to be enhanced to highlight the difference between grayscale values ​​with higher cavitation intensity and the surrounding pixels. The above formula is obtained by Normalization is used to enhance the difference between the mapping optimization factors between the grayscale values ​​with higher cavitation intensity and the grayscale values ​​with lower cavitation intensity, thereby ensuring that the optimized grayscale value mapping can accurately reflect the local grayscale value difference information of the cavitation defect pixels in the image.

[0085] After obtaining the mapping optimization factor of the gray value, the mapping optimization factor can be used to evaluate in the mapping evaluation process to obtain the first The mapping target gray value of gray value :

[0086]

[0087] Formula explanation:

[0088] :Indicates the first The mapping optimization factor of the gray value.

[0089] : No. The cumulative distribution function value corresponding to the gray value.

[0090] : Indicates the minimum value of the number of pixels corresponding to the grayscale value in the image sub-block, that is, the minimum value of the cumulative distribution function.

[0091] : Respectively represent the length and width of the image sub-block, and the unit is the number of pixels of the length and width respectively. In the present invention, .

[0092] : Indicates the number of grayscale values ​​in the image sub-block.

[0093] In the above formula, through The gray value mapping optimization factor optimizes the gray value mapping evaluation process of CLAHE. The higher the mapping optimization factor, the better the gray value mapping evaluation process. The closer the mapping target grayscale value of a grayscale value is to other grayscale values ​​with higher cavitation intensity, the closer the difference is to other grayscale values ​​with higher cavitation intensity.

[0094] Specifically, in the above formula, by As the The weight of the cumulative distribution function of the gray value, so when The higher the value of The mapping effect corresponding to each gray value changes more significantly, thereby preserving the differences between pixels during the mapping process.

[0095] Summary of major steps:

[0096] At this point, the CLAHE equalization process is optimized by the cavitation degree of the pixels in the image, and the gray value mapping result of the CLAHE process is obtained.

[0097] Step S3: The grayscale value of the self-priming pump impeller surface detection image is changed through the grayscale value mapping result of the CLAHE process to complete the image enhancement.

[0098] After obtaining the gray value mapping result of the CLAHE process, the pixel value of each image sub-block can be adjusted according to the gray value mapping result, thereby completing the image sub-block equalization process. After obtaining the mapping of the pixel value of each sub-block of the self-priming pump impeller surface detection image, all image sub-blocks can be integrated through bilinear interpolation to obtain the final image enhancement processing result.

[0099] After obtaining the image enhancement processing results, the enhanced image can be used for subsequent self-priming pump impeller quality inspection. This process can be carried out through neural network inspection, some defect features inspection, or manual inspection. Because the case topic is an image enhancement method, in the end, only the enhanced image needs to be obtained.

[0100] The above technical solutions only reflect the preferred technical solutions of the technical solutions of the present invention. Some changes that may be made to certain parts thereof by technicians in this technical field all reflect the principles of the present invention and fall within the protection scope of the present invention.

Claims

1. An image enhancement method for self-priming pump impeller quality detection, characterized in that: The following steps are involved: Step S1: collecting a surface detection image of a self-priming pump impeller, and dividing the surface detection image of the self-priming pump impeller into image blocks; Step S2: Optimizing the CLAHE equalization process by the cavitation degree of the pixel points in the image to obtain the gray value mapping result of the CLAHE process; The cavitation degree of the pixel point is obtained through the local texture information of the pixel point; The cavitation intensity corresponding to each gray value is evaluated according to the cavitation degree of the pixel point; After obtaining the cavitation intensity of the grayscale histogram in the sub-block, the grayscale value cumulative distribution function of the sub-block is optimized according to the cavitation intensity corresponding to the grayscale value, thereby optimizing the grayscale value mapping and obtaining the grayscale value mapping result of the CLAHE process; Step S3: The grayscale value of the self-priming pump impeller surface detection image is changed through the grayscale value mapping result of the CLAHE process to complete the image enhancement.

2. The image enhancement method for self-priming pump impeller quality detection according to claim 1 is characterized in that: In the step S2: the cavitation degree of the pixel points in the self-priming pump impeller surface detection image is evaluated, and the cavitation degree of the pixel points in the image is evaluated according to the local texture information of the pixel points in the image. After the cavitation degree of the pixel points in the image is obtained, the grayscale histogram evaluation process of each image sub-block can be optimized according to the cavitation degree of the pixel points; In the statistical process of the grayscale histogram, the cavitation intensity corresponding to each grayscale value is evaluated according to the cavitation degree of the pixel points. This cavitation intensity is the degree of cavitation defect pixels in the pixel points corresponding to the grayscale value. The higher the cavitation intensity, the more it is necessary to evaluate the grayscale value mapping result according to the cavitation intensity of each grayscale value in the evaluation of the grayscale value mapping, so as to ensure that the local pixel value difference relationship of the defective pixels in the image will not be eliminated in the process of histogram equalization by limiting the contrast. After obtaining the cavitation intensity of the grayscale histogram in the sub-block, the grayscale value cumulative distribution function of the sub-block is optimized according to the cavitation intensity corresponding to the grayscale value, so as to optimize the grayscale value mapping and obtain the grayscale value mapping result of the CLAHE process.

3. The image enhancement method for self-priming pump impeller quality detection according to claim 1 is characterized in that: The step S2 of obtaining the cavitation degree of the pixel point through the local texture information of the pixel point is as follows: after obtaining the image sub-block of the self-priming pump impeller surface detection image, firstly, the cavitation degree of each pixel point in the image is evaluated. Because when the cavitation defect occurs on the surface of the self-priming pump impeller, it will appear as irregular tiny pits in the image, so the cavitation degree of the central pixel point is evaluated by setting the 24 neighborhoods of the pixel point in the image. Cavitation degree of each pixel: ； Explanation of formula symbols: :Indicates the first The local window of the pixel The gray value of each pixel; :Indicates the first The local window of the pixel Pixels in Grayscale values ​​of adjacent pixels on the right side of the axis; :Indicates the first The local window of the pixel Pixels in Grayscale values ​​of adjacent pixels on the upper side of the axis; :Indicates the first The number of pixels in a local window of pixels; :Indicates the first The local window of the pixel The gradient direction of each pixel; :Indicates the first The local window of the pixel The frequency with which the gradient direction of a pixel appears in the gradient directions of all pixels in the window; : represents the linear normalization function.

4. The image enhancement method for self-priming pump impeller quality detection according to claim 2 is characterized in that: The cavitation intensity corresponding to each gray value is evaluated according to the cavitation degree of the pixel point as follows: After obtaining the cavitation degree of each pixel in the image, the overall cavitation intensity of all pixels corresponding to each gray value can be evaluated according to the cavitation degree of the pixel in the image sub-block. In the image sub-block, when the overall cavitation degree of the pixel corresponding to a gray value is higher, the cavitation intensity of the gray value is also higher. Therefore, in the subsequent gray value mapping process, the gray value mapping process can be optimized by the cavitation intensity of the gray value, and the difference relationship between the gray values ​​with higher cavitation intensity is retained. Gray value of cavitation intensity : ； Explanation of formula symbols: : represents the first The number of pixels corresponding to a gray value.

5. The image enhancement method for self-priming pump impeller quality detection according to claim 2, characterized in that: The step S2 optimizes the gray value cumulative distribution function by the cavitation intensity of the pixel value in the sub-block, and obtains the optimized gray value mapping result according to the optimized cumulative distribution function, specifically: After obtaining the cavitation intensity of each gray value in the image sub-block, the mapping optimization factor corresponding to the gray value can be obtained through the cavitation intensity relationship between the gray values, so that the original gray value is optimized in the gray value mapping evaluation process through the cumulative distribution function through the gray value mapping optimization factor. In the mapping evaluation process, the difference between the gray values ​​with higher cavitation intensity is retained, that is, it is avoided to map several gray values ​​with higher cavitation intensity to gray values ​​with smaller differences. For the first gray value in the image, the gray value mapping optimization factor is used to optimize the original gray value through the cumulative distribution function. Gray value mapping optimization factor : ； In the above formula, the cavitation intensity of the gray value is of Normalization is used to evaluate the mapping optimization factor. In the process of gray value mapping, because the relative relationship between pixels needs to be retained, the mapping optimization factor needs to be evaluated through the difference information of all gray values, so as to ensure that the difference between gray values ​​with large differences is retained in the subsequent gray value mapping process. In addition, because the difference in pixel values ​​between gray values ​​with higher cavitation intensity needs to be retained in the process of regulating the cumulative distribution function by the mapping optimization factor, the difference in pixel values ​​between gray values ​​with higher cavitation intensity needs to be enhanced. Therefore, the difference between the gray values ​​with higher cavitation intensity and the gray values ​​with lower cavitation intensity needs to be enhanced, so as to highlight the difference between the gray values ​​with higher cavitation intensity and the surrounding pixels. In the above formula, Normalization is used to enhance the difference between the mapping optimization factors between the grayscale values ​​with higher cavitation intensity and the grayscale values ​​with lower cavitation intensity, so as to ensure that the optimized grayscale value mapping can accurately reflect the local grayscale value difference information of the cavitation defect pixel points in the image; After obtaining the mapping optimization factor of the gray value, the mapping optimization factor can be used to evaluate in the mapping evaluation process to obtain the first The mapping target gray value of gray value : ； Formula explanation: :Indicates the first The mapping optimization factor of gray values; : No. The cumulative distribution function value corresponding to the gray value; : Indicates the minimum value of the number of pixels corresponding to the grayscale value in the image sub-block, that is, the minimum value of the cumulative distribution function; : Respectively represent the length and width of the image sub-block, the unit is the number of pixels of the length and width respectively, set ; : Indicates the number of gray values ​​in the image sub-block; In the above formula, through The gray value mapping optimization factor optimizes the gray value mapping evaluation process of CLAHE. The higher the mapping optimization factor, the better the gray value mapping evaluation process. The closer the mapping target gray value of a gray value is to other gray values ​​with higher cavitation intensity; In the above formula, through As the The weight of the cumulative distribution function of the gray value, so when The higher the value of The mapping effect corresponding to each gray value changes more significantly, thereby preserving the differences between pixels during the mapping process.

6. The image enhancement method for self-priming pump impeller quality detection according to claim 1, characterized in that: After obtaining the gray value mapping result of the CLAHE process, the pixel value of each image sub-block can be adjusted according to the gray value mapping result, thereby completing the image sub-block equalization process. After obtaining the mapping of the pixel value of each sub-block of the self-priming pump impeller surface detection image, all image sub-blocks can be integrated by bilinear interpolation to obtain the final image enhancement processing result.

Images