A method for detecting heating quality of insulating tube

Abstract

The present invention relates to the field of image data processing, and the present invention relates to a method for detecting the heating quality of a sleeve insulating tube, the method comprising: collecting an image of the surface of the sleeve insulating tube, and performing semantic segmentation on the image to obtain a sleeve insulating tube area; dividing the sleeve insulating tube area into a plurality of image blocks of the same size; calculating the grayscale specificity of each image block; calculating the blur of each image block; performing enhancement processing on the area other than the image block whose blur is less than an index threshold in the image by an unsharp masking algorithm, and obtaining an enhanced image; determining the quality level of the sleeve insulating tube according to the number of pixels whose grayscale values ​​in the enhanced image are greater than the grayscale threshold, and judging whether to alarm. The present invention reduces the influence of surface dirt in the sleeve insulating tube on the judgment of insulation layer shedding defects.

Description

Technical Field

[0001] The present invention relates to the field of image data processing and more specifically to a method for detecting heating quality of an insulating tube. Background Art

[0002] With the improvement of industrialization level and the continuous growth of electricity demand, the construction and maintenance of power infrastructure is particularly important. The casing (i.e. casing for power cables or protective casing for power lines) is a common protective material in power infrastructure, which is mainly used for the laying and maintenance of power lines. If the casing is exposed to harsh environments such as high temperature, high humidity, and ultraviolet rays for a long time, the insulating material may age and degrade, resulting in the shedding of the insulation layer, which in turn leads to faults such as short circuit and leakage, increasing the risk of electric shock and fire, and seriously threatening personal safety and property safety. Therefore, the detection of insulation layer shedding defects on the surface of the casing is an extremely important link to ensure heating quality, safety and stable operation of equipment.

[0003] The existing technology can identify and detect the defects of insulating materials in advance through heating quality detection based on image processing, prevent electrical accidents caused by insulation failure, and thus improve the safety of the overall power system. For example, the Chinese patent application document with publication number CN108535612A discloses a method for constructing an insulation detection database for SF6 insulating bushings and its detection method and device. The detection database constructed by this scheme contains fitting parameter groups of SF6 insulating bushings with different aging degrees, which can be conducive to further using the database to perform insulation detection on SF6 insulating bushings.

[0004] However, dust and other substances in the external environment may adhere to the surface of the insulating tube to form dirt. The dirt area and the insulation layer peeling area have similar image features. Now it is necessary to detect the potential insulation layer peeling defects on the surface of the insulating tube without detecting the dirt. In order to improve the clarity of the surface image of the insulating tube and enhance the details of the image, the existing technology usually uses an unsharp masking algorithm to enhance the surface image of the insulating tube. At this time, it is necessary to blur each pixel in the surface image of the insulating tube. Blurring each pixel may cause the details of the potential insulation layer to be blurred, which may make it more difficult to distinguish between the dirt area and the insulation layer peeling area, thereby reducing the detection accuracy. Summary of the invention

[0005] In order to solve the technical problem that the above-mentioned dirt area may affect the accuracy of detecting the insulation layer peeling defect area, the present invention provides solutions in the following aspects.

[0006] A method for detecting heating quality of a sleeve insulating tube comprises: collecting an image of the surface of the sleeve insulating tube, and performing semantic segmentation on the image to obtain a sleeve insulating tube region; dividing the sleeve insulating tube region into a plurality of image blocks of the same size; calculating the grayscale specificity of each image block, wherein the grayscale specificity of any image block is inversely proportional to the mean grayscale value of the pixel points in the image block, inversely proportional to the mean grayscale value of all pixel points in the region other than the image block in the sleeve insulating tube region, and inversely proportional to the mean grayscale value of all pixel points in the sleeve insulating tube region; calculating the grayscale specificity of each image block; and calculating the grayscale specificity of each image block. k The blurriness of the image blocks B k : , for the k Image blocks: G k is the grayscale specificity of the image block, DC k is the extreme difference of the grayscale values ​​of all pixels in the image block, J k is the number of pixels that are corner points in the image block, J k,r The image block is the first image block in the insulating tube area. r The number of pixels that are corner points in an image block, exp() is e The exponential function with base , R is the number of all image blocks, j , k , r are all positive integers; the areas other than the image blocks whose blurriness in the image is less than the exponential threshold are enhanced by the unsharp masking algorithm to obtain an enhanced image, wherein or is the enhancement factor of the unsharp masking algorithm, and or <0; determine the quality level of the insulating tube according to the number of pixels in the enhanced image whose grayscale value is greater than the grayscale threshold, and determine whether to alarm.

[0007] The beneficial effects are as follows: the present invention retains the key features of potential defects by selectively blurring the image, while avoiding loss of details, thereby effectively distinguishing between dirt and insulation layer shedding defects, improving detection accuracy and sensitivity, avoiding misjudgment, and improving detection efficiency and robustness. The present invention obtains the blurriness of each image block in the surface image of the insulating tube by analyzing the surface image features of the insulating tube, and then screens the image blocks that do not need blurring, so as to improve the accuracy of subsequent detection of potential insulation layer shedding defects on the surface of the insulating tube.

[0008] Preferably, calculate k Grayscale specificity of image blocks G k The formula is:

[0009] ;

[0010] in D is the mean grayscale value of all pixels in the insulating tube area, Yes k For the k The mean gray value of all pixels in an image block, DE k The image of the surface of the insulating tube is as follows: k The mean grayscale value of all pixels in the area outside the image block, norm () is the normalization function.

[0011] The present invention reflects the possibility that the pixels included in the image block belong to the insulating layer detachment defect area by calculating the grayscale specificity of the image block (the outlier reflects the difference between the pixels in the image block and the pixels of the overall image), so that the fuzziness subsequently calculated by the grayscale specificity can reflect the possibility of the pixels in the image block including the insulating layer detachment defect area.

[0012] Preferably, a method for detecting heating quality of a sleeve insulating tube further comprises: before calculating the fuzziness of each image block, correcting the grayscale specificity of each image block, and using the corrected grayscale specificity of each image block as its grayscale specificity at the current moment; wherein, for the first k The grayscale specificity of each image block is corrected by: , For the k The updated grayscale specificity of the image block, S k For the k The number of pixels in an image block, L k For the k The number of all gray levels corresponding to all pixels in an image block.

[0013] Considering that after the insulation layer falls off, the exposed metal or other material cracks, scratches or particles cause the local texture of the insulation tube surface to become complex, and the grayscale value of the corresponding image pixel is no longer smooth and uniform. Therefore, the present invention uses a grayscale value representing the smoothness of the image block. The grayscale specificity of the image block is corrected so that the grayscale specificity of each image block can better reflect the difference between the grayscale value of the pixel point in the image block and the pixel point of the entire image.

[0014] Preferably, the region other than the image block whose blurriness is less than the exponential threshold in the image is enhanced by an unsharp masking algorithm to obtain an enhanced image, comprising: taking the region other than the image block whose blurriness is less than the exponential threshold as the target region image, blurring the target region image by a Gaussian blur algorithm to obtain a blurred image; subtracting the blurred image from the target region image to obtain a detail image; obtaining the detail image and the enhancement factor or The product of is added to the target area image to obtain an enhanced image.

[0015] Preferably, subtracting the blurred image from the target area image to obtain a detailed image includes: taking any pixel point in the target area image as a first pixel point, and taking the coordinates of the first pixel point as a first coordinate, subtracting the grayscale value of the pixel point of the blurred image at the first coordinate from the grayscale value of the first pixel point to obtain the grayscale value of the detailed image at the first coordinate; taking the coordinates of all pixel points in the target area image as a first coordinate set, determining the grayscale values ​​of the pixel points of the detailed image at each coordinate in the first coordinate set to obtain an enhanced image.

[0016] Preferably, the detail image and the enhancement factor are obtained or The product of the product and the target area image are added to obtain the enhanced image, including: taking any pixel point in the detail image as the second pixel point, and taking the coordinate of the second pixel point as the second coordinate, multiplying the gray value of the second pixel point by the enhancement factor or Then, the grayscale value of the pixel point of the blurred image at the first coordinate is added to obtain the grayscale value of the detail image at the first coordinate; the coordinates of all pixel points in the detail image are taken as the second coordinate set, and the grayscale values ​​of the pixel points of the enhanced image at each coordinate in the second coordinate set are determined to obtain the enhanced image.

[0017] Preferably, the image is semantically segmented by a semantic segmentation model, and constructing the semantic segmentation model includes: collecting a training set consisting of multiple training images, the pixel points in the insulating tube area of ​​the training image are marked as 1, and the pixel points in the area outside the insulating tube area of ​​the training image are marked as 0; constructing a deep convolutional neural network model; training the deep convolutional neural network model by using the training set to obtain the semantic segmentation model.

[0018] Preferably, a method for detecting the heating quality of an insulating tube also includes graying the image to obtain the grayscale value of each pixel in the image: performing noise reduction and filtering on the image; for any pixel in the image after noise reduction and filtering, weightedly summing the R channel color value, G channel color value, and B channel color value of the pixel to obtain the grayscale value of the pixel.

[0019] The present invention calculates grayscale values ​​by noise reduction, filtering, and weighted summation of RGB channel color values, which can effectively improve image quality, reduce noise interference, enhance the visual consistency of the image, and provide more reliable and accurate input for subsequent image analysis and processing.

[0020] Preferably, the quality level of the insulating tube sleeve is determined according to the number of pixels in the enhanced image whose grayscale values ​​are greater than the grayscale threshold, and it is judged whether to alarm, including: in response to the number of pixels in the enhanced image whose grayscale values ​​are greater than the grayscale threshold being less than a first threshold, determining that the insulating tube sleeve corresponding to the image is of the first quality level, and marking the insulating tube sleeve corresponding to the image as qualified; in response to the number of pixels in the enhanced image whose grayscale values ​​are greater than the grayscale threshold being greater than or equal to the first threshold and less than the second threshold, determining that the insulating tube sleeve corresponding to the image is of the second quality level, and marking the insulating tube sleeve corresponding to the image as to be inspected manually; in response to the number of pixels in the enhanced image whose grayscale values ​​are greater than the grayscale threshold being greater than or equal to the second threshold, determining that the insulating tube sleeve corresponding to the image is of the third quality level, and alarming.

[0021] The present invention evaluates the quality of the insulating sleeve according to the number of pixel points whose grayscale values ​​are greater than the grayscale threshold in the enhanced image, and adopts different processing methods for different quality levels to optimize the production process and save costs.

[0022] The beneficial effects of the present invention are:

[0023] The present invention can reduce the influence of surface dirt in the insulating tube on the judgment of insulation layer shedding defects by analyzing the surface image features of the insulating tube and determining the blurriness of each image block, thereby improving the accuracy of quality inspection of the insulating tube. By setting a threshold and screening image blocks that do not need blurring, key features can be ensured to remain clear. This selective blurring helps to retain the subtle features of the insulation layer shedding defect in the potential image and improve the accuracy of detection. BRIEF DESCRIPTION OF THE DRAWINGS

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

[0025] Figure 1 The present invention is a flowchart of a method for detecting heating quality of an insulating tube according to an embodiment of the present invention. DETAILED DESCRIPTION

[0026] The following will be combined with the drawings in the embodiments of the present invention to clearly and completely describe the technical solutions in the embodiments of the present invention. Obviously, the described embodiments are part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative work are within the scope of protection of the present invention.

[0027] The specific embodiments of the present invention are described in detail below with reference to the accompanying drawings.

[0028] Figure 1 The present invention is a flowchart of a method for detecting heating quality of an insulating tube according to an embodiment of the present invention.

[0029] like Figure 1 As shown, a method for detecting heating quality of an insulating tube includes steps S1 to S6.

[0030] Step S1: collecting an image of the surface of the insulating tube, and performing semantic segmentation on the image to obtain an insulating tube region.

[0031] It should be noted that the semantic segmentation model is used to classify each pixel in the grayscale image. In the present invention, the semantic segmentation model divides the pixels in the grayscale image into two categories, one is the pixels in the insulating tube area, and the other is the pixels that do not belong to the insulating tube area (or called the background area). In one embodiment, the present invention performs semantic segmentation on the image through a semantic segmentation model, wherein constructing the semantic segmentation model includes: collecting a training set consisting of multiple training images, the pixels in the insulating tube area of ​​the training image are marked as 1, and the pixels in the area outside the insulating tube area of ​​the training image are marked as 0; constructing a deep convolutional neural network model; training the deep convolutional neural network model through the training set to obtain the semantic segmentation model.

[0032] In one embodiment, the present invention further comprises graying the image to obtain the gray value of each pixel in the image: performing noise reduction and filtering on the image; for any pixel in the image after noise reduction and filtering, weighted summing the R channel color value, the G channel color value, and the B channel color value of the pixel to obtain the gray value of the pixel; determining the gray value of each pixel in the image to obtain a grayscale image of the image. In one embodiment, for any pixel, the formula for calculating its gray value is:

[0033] Gray=r×0.299+g×0.587+b×0.114.

[0034] Gray is the grayscale value of the pixel, r is the R channel color value of the pixel, g is the G channel color value of the pixel, and b is the B channel color value of the pixel.

[0035] Step S2: Divide the insulating tube area into a plurality of image blocks of the same size.

[0036] In one embodiment, the insulating tube region is divided into a plurality of 5×5 image blocks, and any image block includes 25 pixels.

[0037] In another embodiment, since the edge shape of the insulating tube region is irregular, it is impossible to divide all the pixels of the insulating tube region into a plurality of image blocks of the same size. Therefore, the image of the surface of the insulating tube is divided into a plurality of 5×5 initial image blocks, and the image blocks including at least Q The initial image block of pixels is the image block used in subsequent calculations. Q It is a positive integer such as 2, 3, 4, etc. that is less than 25 (the total number of pixels in the image block).

[0038] Step S3: Calculate the grayscale specificity of each image block.

[0039] Among them, the grayscale specificity of any image block is inversely proportional to the mean grayscale value of the pixels in the image block, inversely proportional to the mean grayscale value of all pixels in the insulating tube area except the image block, and inversely proportional to the mean grayscale value of all pixels in the insulating tube area.

[0040] In an embodiment of the present invention, the calculation k Grayscale specificity of image blocks G k The formula is: .in D is the mean grayscale value of all pixels in the insulating tube area, Yes k For the k The mean gray value of all pixels in an image block, DE k The image of the surface of the insulating tube is as follows: k The mean grayscale value of all pixels in the area outside the image block, norm () is the normalization function.

[0041] It should be noted that the grayscale value of the area where the insulation layer of the insulating tube falls off is usually smaller. This is because after the insulation layer of the insulating tube falls off, the exposed metal or other materials usually have a higher reflectivity, resulting in a lower grayscale value in the image. The grayscale value of the pixel corresponding to the insulation layer in the normal area is larger. Therefore, in the grayscale image, the detached area is usually displayed as a darker part. Therefore, when analyzing the grayscale specificity of each image block, the lower the grayscale value of the pixel in each image block, the more likely it is that this image block belongs to or includes the pixel in the area where the insulation layer falls off.

[0042] Based on this, the formula Yes k The smaller the k The smaller the mean gray value of all pixels in an image block is, the more likely it is a potential insulation layer peeling defect, and the greater its grayscale specificity. It indicates the k The relative gray value of the image block in the entire insulating tube image. The larger the value, the more k After the image block, the grayscale mean of the whole image will become larger, which can indirectly reflect the k The smaller the relative gray value of the image block in the entire insulating tube surface image, the smaller the relative gray value of the image block k The smaller the grayscale mean value in an image block, the greater the credibility, the more likely this image block is a potential insulation layer peeling defect, and the greater its grayscale specificity. The smaller the middle, the smaller the k The smaller the grayscale mean value in an image block, the greater the credibility, the more likely this image block is a potential insulation layer detachment defect, and the greater its grayscale specificity.

[0043] Furthermore, before calculating the blurriness of each image block, the grayscale specificity of each image block is corrected, and the corrected grayscale specificity of each image block is used as the grayscale specificity of the current moment; k The grayscale specificity of each image block is corrected by: , For the k The updated grayscale specificity of the image block, S k For the k The number of pixels in an image block, L k For the k The number of all gray levels corresponding to all pixels in an image block.

[0044] It should be noted that after the insulation layer falls off, the exposed metal or other material cracks, scratches or particles cause the local texture of the surface of the insulating tube to become complex, and the grayscale value of the pixels of the corresponding image is no longer smooth and uniform. k The more chaotic the grayscale values ​​of the pixels in an image block are (i.e. The larger the value), the greater the k The greater the possibility that there is a potential insulation layer peeling defect in an image block, the greater the grayscale specificity of the image block will be.

[0045] Step S4: Calculate the blur of each image block.

[0046] Among them, calculate k The blurriness of the image blocks B k :

[0047] .

[0048] For k Image blocks: G k is the grayscale specificity of the image block, DC k is the extreme difference of the grayscale values ​​of all pixels in the image block, J k is the number of pixels that are corner points in the image block, J k,r The image block is the first image block in the insulating tube area. r The number of pixels that are corner points in an image block, exp() is e The exponential function with base , R is the number of all image blocks, j , k , r All are positive integers.

[0049] It should be noted that, since the dirt area that may exist on the surface of the insulating tube also has a lower grayscale value and a more intensive grayscale change, it is necessary to distinguish between the dirt area that may exist on the surface of the insulating tube and the potential insulation layer shedding defect area.

[0050] Considering that the dirt area has more blurred boundaries and more complex textures than the insulation layer detachment defect area. When analyzing the blurriness of each image block, the smaller the grayscale range and the more corner points in each image block, the more likely it is that this image block belongs to the dirt area, and this block can be blurred normally in the subsequent blurring process. On the contrary, if the grayscale range in an image block is larger and the number of corner points is smaller, it can be said that the possibility that this image block belongs to the insulation layer detachment defect area is greater, then this image block is less likely to be blurred in the subsequent blurring process in order to retain the detailed information in the block as much as possible, and the corresponding blurriness is lower.

[0051] Step S5: enhancing the area other than the image block whose blurriness is less than the index threshold in the image by using an unsharp masking algorithm to obtain an enhanced image.

[0052] in or is the enhancement factor of the unsharp masking algorithm, and or <0. In one embodiment, or The experience value is -1.1.

[0053] In one embodiment, or The method comprises the following steps: performing enhancement processing on the area outside the image block whose blurriness is less than the exponential threshold in the image by using an unsharp masking algorithm as an enhancement factor, and obtaining an enhanced image, including: taking the area outside the image block whose blurriness is less than the exponential threshold as a target area image, blurring the target area image by using a Gaussian blur algorithm to obtain a blurred image; subtracting the blurred image from the target area image to obtain a detail image; obtaining the detail image and the enhancement factor or In this embodiment, when blurring the target area image by using the Gaussian blur algorithm, the blur radius is set to 3 and the standard deviation is set to 1.

[0054] Among them, subtracting the blurred image from the target area image to obtain a detailed image includes: taking any pixel point in the target area image as the first pixel point, and taking the coordinates of the first pixel point as the first coordinate, subtracting the grayscale value of the pixel point of the blurred image at the first coordinate from the grayscale value of the first pixel point to obtain the grayscale value of the detailed image at the first coordinate; taking the coordinates of all pixel points in the target area image as the first coordinate set, determining the grayscale values ​​of the pixel points of the detailed image at each coordinate in the first coordinate set to obtain an enhanced image.

[0055] Furthermore, the detail image and the enhancement factor are obtained orThe product of the product and the target area image are added to obtain the enhanced image, including: taking any pixel point in the detail image as the second pixel point, and taking the coordinate of the second pixel point as the second coordinate, multiplying the gray value of the second pixel point by the enhancement factor or Then, the grayscale value of the pixel point of the blurred image at the first coordinate is added to obtain the grayscale value of the detail image at the first coordinate; the coordinates of all pixel points in the detail image are taken as the second coordinate set, and the grayscale values ​​of the pixel points of the enhanced image at each coordinate in the second coordinate set are determined to obtain the enhanced image.

[0056] It should be noted that in image enhancement, detail images usually represent the high-frequency parts of the image (details, edges, etc.). or is negative (e.g. or =-1.1), the pixel values ​​in the detail image will be inverted, thereby suppressing the details in the original image and producing a blurring effect. In contrast, the image that does not belong to the target area image in the insulating tube image will be emphasized, that is, the image that does not belong to the target area image will be clearer than the target area image.

[0057] Step S6: Determine the quality level of the insulating tube according to the number of pixels in the enhanced image whose grayscale values ​​are greater than the grayscale threshold, and determine whether to issue an alarm.

[0058] In one embodiment, determining the quality level of the insulating tube according to the number of pixels in the enhanced image whose grayscale values ​​are greater than the grayscale threshold, and determining whether to alarm, includes:

[0059] In response to the number of pixels in the enhanced image whose grayscale values ​​are greater than the grayscale threshold being less than the first threshold, the insulating tube corresponding to the image is determined to be of the first quality level, and the insulating tube corresponding to the image is marked as qualified. For the insulating tube marked as qualified, manual inspection and maintenance are not required.

[0060] In response to the number of pixels in the enhanced image whose grayscale values ​​are greater than the grayscale threshold being greater than or equal to the first threshold and less than the second threshold, the insulating tube corresponding to the image is determined to be of the second quality level, and the insulating tube corresponding to the image is marked as to be inspected manually. For the insulating tubes marked as to be inspected manually, relevant personnel are required to uniformly inspect the insulating tubes of this type at a predetermined inspection cycle, for example, the insulating tubes marked as to be inspected manually are inspected on Wednesdays of each week to determine whether the insulating tubes of this type have the defect of insulation layer shedding.

[0061] In response to the number of pixels in the enhanced image whose grayscale values ​​are greater than the grayscale threshold being greater than or equal to the second threshold, it is determined that the set of insulating pipes corresponding to the image is of the third quality level, and an alarm is issued to notify relevant personnel to go to the set of insulating pipes at the current moment to inspect and repair them.

[0062] In one embodiment, the first threshold is 0.002 times the number of pixels in the insulating tube area, and the second threshold is 0.01 times the number of pixels in the insulating tube area.

[0063] In the description of this specification, "plurality" or "several" means at least two, such as two, three or more, etc., unless otherwise clearly and specifically defined.

[0064] Although this specification has shown and described a number of embodiments of the present invention, it will be apparent to those skilled in the art that such embodiments are provided by way of example only. Those skilled in the art will conceive of many modifications, changes and alternatives without departing from the ideas and spirit of the present invention. It should be understood that in the practice of the present invention, various alternatives to the embodiments of the present invention described herein may be employed.

Claims

1. A method for detecting the heating quality of an insulating tube, characterized in that: include: Collecting an image of the surface of the insulating tube, and performing semantic segmentation on the image to obtain an insulating tube region; Dividing the insulating tube area into a plurality of image blocks of the same size; Calculate the grayscale specificity of each image block, the grayscale specificity of any image block is inversely proportional to the average grayscale value of the pixels in the image block, inversely proportional to the average grayscale value of all pixels in the area other than the image block in the insulating tube area, and inversely proportional to the average grayscale value of all pixels in the insulating tube area; Calculate the k The blurriness of the image blocks B k : , for the k Image blocks: G k is the grayscale specificity of the image block, DC k is the extreme difference of the grayscale values ​​of all pixels in the image block, J k is the number of pixels that are corner points in the image block, J k,r The image block is the first image block in the insulating tube area. r The number of pixels that are corner points in an image block, exp() is e The exponential function with base , R is the number of all image blocks, j , k , r All are positive integers; The method further comprises: performing enhancement processing on the area other than the image block whose blurriness is less than the index threshold in the image by using an unsharp masking algorithm, and obtaining an enhanced image, including: Taking the area outside the image block whose blurriness is less than the index threshold as the target area image, blurring the target area image by using the Gaussian blur algorithm to obtain a blurred image; Subtracting the blurred image from the target area image to obtain a detail image; Obtain the detail image and enhancement factor η The product of is added to the target area image to obtain an enhanced image; wherein η is the enhancement factor of the unsharp masking algorithm, and η <0; determine the quality level of the insulating tube according to the number of pixels in the enhanced image whose grayscale value is greater than the grayscale threshold, and determine whether to alarm.

2. A method for detecting heating quality of a sleeve insulating tube according to claim 1, characterized in that: Calculate the k Grayscale specificity of image blocks G k The formula is: ； in D is the mean grayscale value of all pixels in the insulating tube area, DA k For the k The mean gray value of all pixels in an image block, DE k The image of the surface of the insulating tube is as follows: k The mean grayscale value of all pixels in the area outside the image block, norm () is the normalization function.

3. A method for detecting heating quality of a sleeve insulation tube according to claim 1, characterized in that: Also includes: Before calculating the blurriness of each image block, the grayscale specificity of each image block is corrected, and the corrected grayscale specificity of each image block is used as the grayscale specificity of the image block at the current moment; Among them, k The grayscale specificity of each image block is corrected by: , For the k The updated grayscale specificity of the image block, S k For the k The number of pixels in an image block, L k For the k The number of all gray levels corresponding to all pixels in an image block.

4. A method for detecting heating quality of a sleeve insulation tube according to claim 1, characterized in that: Subtracting the blurred image from the target area image to obtain a detail image comprises: Taking any pixel point in the target area image as the first pixel point and the coordinate of the first pixel point as the first coordinate, subtracting the grayscale value of the pixel point of the blurred image at the first coordinate from the grayscale value of the first pixel point to obtain the grayscale value of the detail image at the first coordinate; The coordinates of all pixels in the target area image are taken as a first coordinate set, and the grayscale values ​​of the pixels in the detail image at each coordinate in the first coordinate set are determined to obtain an enhanced image.

5. The method for detecting heating quality of a sleeve insulating tube according to claim 1, characterized in that: Obtain the detail image and enhancement factor η The product of , adding the product to the target area image to obtain an enhanced image comprises: Take any pixel in the detail image as the second pixel, and take the coordinate of the second pixel as the second coordinate, and multiply the gray value of the second pixel by the enhancement factor η Then add the gray value of the pixel point of the blurred image at the first coordinate to obtain the gray value of the detail image at the first coordinate; The coordinates of all pixels in the detail image are taken as the second coordinate set, and the grayscale values ​​of the pixels of the enhanced image at each coordinate in the second coordinate set are determined to obtain the enhanced image.

6. A method for detecting heating quality of a sleeve insulating tube according to claim 1, characterized in that: Performing semantic segmentation on the image by using a semantic segmentation model, constructing the semantic segmentation model includes: Collecting a training set consisting of a plurality of training images, wherein the pixel points in the insulating tube region of the training images are marked as 1, and the pixel points in the region other than the insulating tube region of the training images are marked as 0; Constructing a deep convolutional neural network model, wherein the deep convolutional neural network model is a neural network model of a U-Net architecture; The deep convolutional neural network model is trained by the training set, and the semantic segmentation model is obtained when the deep convolutional neural network model is trained a predetermined number of times by the training set.

7. A method for detecting heating quality of a sleeve insulating tube according to claim 1, characterized in that: The method further includes graying the image to obtain the gray value of each pixel in the image: Performing noise reduction and filtering processing on the image; For any pixel in the image after noise reduction and filtering, the R channel color value, the G channel color value, and the B channel color value of the pixel are weightedly summed to obtain the grayscale value of the pixel.

8. A method for detecting heating quality of a sheathed insulating tube according to claim 1, characterized in that: The quality level of the insulating tube is determined based on the number of pixels in the enhanced image whose grayscale value is greater than the grayscale threshold, and whether an alarm is issued is determined, including: In response to the number of pixels in the enhanced image whose grayscale values ​​are greater than the grayscale threshold being less than a first threshold, the sleeve insulating tube corresponding to the image is determined to be of the first quality level, and the sleeve insulating tube corresponding to the image is marked as qualified; in response to the number of pixels in the enhanced image whose grayscale values ​​are greater than the grayscale threshold being greater than or equal to the first threshold and less than the second threshold, the sleeve insulating tube corresponding to the image is determined to be of the second quality level, and the sleeve insulating tube corresponding to the image is marked as awaiting manual inspection; in response to the number of pixels in the enhanced image whose grayscale values ​​are greater than the grayscale threshold being greater than or equal to the second threshold, the sleeve insulating tube corresponding to the image is determined to be of the third quality level, and an alarm is issued; wherein the first threshold is k1 times the total number of pixels in the sleeve insulating tube area, the second threshold is k2 times the number of pixels in the sleeve insulating tube area, k1<1, k2<1, and k2<1, <k1。

Images