Method and system for automatically analyzing particle size of powder coating

Abstract

The invention relates to a powder coating particle size automatic analysis method, which comprises the following steps: obtaining a powder coating particle image shot by an electron microscope, carrying out feature extraction on the particle image to obtain a plurality of feature maps, dividing each feature map into image blocks with the same size, obtaining areas of the image blocks on the particle image, obtaining a foreground score of the image block by using the average gray value and the contour line length of the region; selecting a plurality of target image blocks from each feature map according to the foreground score and particle size distribution information in the particle image, calculating the attention of the target image blocks at each layer of the encoder, and then calculating the attention of the target image blocks and the filtered image blocks based on the average gray value; and obtaining an output corresponding to the target image block in the encoder, taking the output as an input of a decoder, obtaining a particle frame in the particle image through the decoder, and obtaining particle size information of the powder coating according to the particle frame.

Description

Technical Field

[0001] The invention relates to the field of coatings, and in particular to an automatic analysis method and system for particle size of powder coatings. Background Art

[0002] Powder coating is a type of coating in the form of solid powder. Compared with traditional liquid coatings, powder coatings produce almost no volatile organic compounds (VOCs) during the construction process, which meets the requirements of modern industry for green environmental protection. Its main components include resins, curing agents, pigments, fillers and additives, which together give the coating excellent mechanical strength, chemical corrosion resistance and beautiful decorative effects. The particle size of powder coating directly affects the spraying quality and the final coating performance. Too wide a particle size distribution may lead to reduced spraying efficiency and uneven coating thickness; too small a particle size may cause excessive waste. Therefore, accurate analysis and statistics of the particle size of powder coating particles are of great significance for optimizing production processes and improving product performance. However, due to the complex morphology of powder coating particles, the wide range of particle size distribution and the possibility of adhesion between particles, traditional particle size analysis methods such as laser particle size analysis have certain limitations in accuracy and efficiency. Electron microscope images have high resolution and clear details, and can capture the microscopic morphology of powder particles. In the past, the particle size of powder coatings analyzed in electron microscope images was mostly measured manually, such as using tools such as Nano Measurer. However, on the one hand, this method requires a lot of manual labor. On the other hand, there is a certain error in pulling the particle size with the mouse, and the boundary cannot be accurately determined. Moreover, since the particles are not perfectly round, the results obtained by pulling the particle size from different angles are also different. Summary of the invention

[0003] In order to automatically identify powder coating particles and analyze the particle size from multiple angles, firstly, the present invention provides a method for automatically analyzing the particle size of powder coatings, the method comprising the following steps: Obtaining a powder coating particle image photographed by an electron microscope, performing feature extraction on the particle image to obtain a plurality of feature maps, dividing each feature map into image blocks of the same size, obtaining an area of ​​the image block on the particle image, and obtaining a foreground score of the image block using an average gray value and a contour line length of the area; Select multiple target image blocks from each feature map according to the foreground score and the particle size distribution information in the particle image. At each layer of the encoder, the attention of the target image block is first calculated, and then the attention of the target image block and the image block filtered based on the average gray value is calculated. The output corresponding to the target image block in the encoder is obtained and used as the input of the decoder. The particle frame in the particle image is obtained through the decoder, and the particle size information of the powder coating is obtained according to the particle frame.

[0004] Preferably, the foreground score of the image block is obtained by using the average gray value and the contour length of the region, specifically: The average grayscale value is divided by 255 to obtain the normalized grayscale value, and the ratio of the contour length of the area to the perimeter of the area is calculated; The weighted sum of the normalized gray value and the ratio is taken as the foreground score.

[0005] Preferably, the step of selecting a plurality of target image blocks from each feature image according to the foreground score and the particle size distribution information in the particle image is as follows: The connected component labeling method is used to calculate the area of ​​each white region in the binary particle image; Cluster the white areas according to their areas. The number of clusters in the cluster is the same as the number of feature maps. Calculate the average area of ​​the white areas in each cluster. Establish the correspondence between the feature maps and clusters based on the average area. The larger the feature map size, the smaller the average area of ​​the corresponding cluster. The particle size distribution information in the particle image is obtained by calculating the ratio of the number of white areas in the cluster corresponding to the feature map; The number of target image blocks of each feature map is obtained according to the particle size distribution information in the particle image and the preset number of target image blocks, and multiple image blocks with the largest foreground scores are selected from the image blocks of each feature map as the target image blocks of the feature map, wherein the number of the multiple image blocks is equal to the number of target image blocks of the feature map.

[0006] Preferably, the calculation of the attention of the target image block and the image block filtered based on the average gray value is specifically as follows: The average gray value of the area that does not contain powder coating particles in the training sample is used as the threshold. If the average gray value of the image block is less than the threshold, the image block is filtered out; The target image patch is used to get the query, and the filtered image patch is used to get the key and value, and the attention is calculated.

[0007] Preferably, the particle size information of the powder coating is obtained according to the particle frame, specifically: Obtaining a vertical central axis of the particle frame, rotating the central axis according to a preset step length, and counting the length of the central axis in the particle frame after each rotation where the grayscale value is greater than the average grayscale value of the particle frame; After the rotation is completed, the maximum length is taken as the length of the powder coating particle, and the minimum length is taken as the width of the powder coating particle; The length and width of all particles in the particle image are counted to obtain the length distribution histogram and width distribution histogram.

[0008] Secondly, the present invention also provides a powder coating particle size automatic analysis system, the system comprising the following modules: A feature extraction module is used to obtain a powder coating particle image taken by an electron microscope, perform feature extraction on the particle image to obtain a plurality of feature maps, divide each feature map into image blocks of the same size, obtain an area of ​​the image block on the particle image, and obtain a foreground score of the image block using an average gray value and a contour line length of the area; An encoding module is used to select multiple target image blocks from each feature map according to the foreground score and the particle size distribution information in the particle image, and first calculate the attention of the target image block at each layer of the encoder, and then calculate the attention of the target image block and the image block filtered based on the average gray value; The decoding and analysis module is used to obtain the output corresponding to the target image block in the encoder and use it as the input of the decoder. The particle frame in the particle image is obtained through the decoder, and the particle size information of the powder coating is obtained according to the particle frame.

[0009] Preferably, the foreground score of the image block is obtained by using the average gray value and the contour length of the region, specifically: The average grayscale value is divided by 255 to obtain the normalized grayscale value, and the ratio of the contour length of the area to the perimeter of the area is calculated; The weighted sum of the normalized gray value and the ratio is taken as the foreground score.

[0010] Preferably, the step of selecting a plurality of target image blocks from each feature image according to the foreground score and the particle size distribution information in the particle image is as follows: The connected component labeling method is used to calculate the area of ​​each white region in the binary particle image; Cluster the white areas according to their areas. The number of clusters in the cluster is the same as the number of feature maps. Calculate the average area of ​​the white areas in each cluster. Establish the correspondence between the feature maps and clusters based on the average area. The larger the feature map size, the smaller the average area of ​​the corresponding cluster. The particle size distribution information in the particle image is obtained by calculating the ratio of the number of white areas in the cluster corresponding to the feature map; The number of target image blocks of each feature map is obtained according to the particle size distribution information in the particle image and the preset number of target image blocks, and multiple image blocks with the largest foreground scores are selected from the image blocks of each feature map as the target image blocks of the feature map, wherein the number of the multiple image blocks is equal to the number of target image blocks of the feature map.

[0011] Preferably, the calculation of the attention of the target image block and the image block filtered based on the average gray value is specifically as follows: The average gray value of the area that does not contain powder coating particles in the training sample is used as the threshold. If the average gray value of the image block is less than the threshold, the image block is filtered out; The target image patch is used to get the query, and the filtered image patch is used to get the key and value, and the attention is calculated.

[0012] Preferably, the particle size information of the powder coating is obtained according to the particle frame, specifically: Obtaining a vertical central axis of the particle frame, rotating the central axis according to a preset step length, and counting the length of the central axis in the particle frame after each rotation where the grayscale value is greater than the average grayscale value of the particle frame; After the rotation is completed, the maximum length is taken as the length of the powder coating particle, and the minimum length is taken as the width of the powder coating particle; The length and width of all particles in the particle image are counted to obtain the length distribution histogram and width distribution histogram.

[0013] Finally, the present invention also provides a computer-readable storage medium, on which a computer program is stored, and when the computer program is executed by a processor, the method described in the first aspect is implemented. In addition, the present invention also provides a computer program, which includes the module of the system described in the second aspect.

[0014] The present invention uses an automated method to identify the particle size of powder coatings. The foreground score of the image block on the feature map is calculated by the average gray value and contour line length of the image block on the feature map in the area of ​​the particle image. The image block containing particles is selected from all the image blocks on all the feature maps. The image blocks with an average gray value less than a threshold are filtered to reduce the interference of the image blocks without particles on the image blocks with a larger foreground score, thereby improving the accuracy of identification. In addition, the present invention uses a rotating central axis to calculate the length and width of each particle frame, which can adapt to the irregularity of the particle shape and improve the accuracy of analysis. BRIEF DESCRIPTION OF THE DRAWINGS

[0015] Figure 1 This is a flow chart of Embodiment 1; Figure 2 This is a picture of powder coating particles taken by electron microscope; Figure 3 Powder coating particles Figure 2 Value results; Figure 4 It is a statistical histogram of the length of particles in the powder particle map. DETAILED DESCRIPTION

[0016] In order to make the purpose, technical solution and advantages of the present invention more clearly understood, the present invention is further described in detail below in conjunction with the accompanying drawings and embodiments. It should be understood that the specific embodiments described herein are only used to explain the present invention and are not intended to limit the present invention.

[0017] It should be noted that, in this article, the terms "include", "comprises" or any other variations thereof are intended to cover non-exclusive inclusion, so that a process, method, article or terminal including a series of elements includes not only those elements, but also other elements not explicitly listed, or also includes elements inherent to such process, method, article or terminal. In the absence of further restrictions, an element defined by the sentence "comprises a ..." does not exclude the presence of other identical elements in the process, method, article or terminal including the element.

[0018] The first implementation is Figure 1 As shown, the following steps are included: S1, obtaining a powder coating particle image photographed by an electron microscope, performing feature extraction on the particle image to obtain a plurality of feature maps, dividing each feature map into image blocks of the same size, obtaining an area of ​​the image block on the particle image, and obtaining a foreground score of the image block using an average gray value and a contour line length of the area; Use an electron microscope (EM) to take images of powder coating particles to obtain high-resolution particle images. Figure 2 As shown, in the particle image, the particle area is brighter, while the background is darker. Feature extraction is performed on the particle image to generate multiple feature maps. Among them, the feature maps have different sizes, and the feature extraction of the particle image is performed using a backbone network. Then the feature map is divided into image blocks of the same size. For example, a 16×16 feature block can be divided into 16 4×4 image blocks. Of course, feature maps of other sizes are also divided into multiple 4×4 image blocks.

[0019] Each image block corresponds to a part of the original particle image. Due to the different receptive fields of different feature maps, the original image may contain particles of different numbers and sizes. The average gray value of each image block in the original image is calculated. The higher the gray value, the higher the brightness of the area, and the more likely it is to contain particles. However, even if the same particles have different brightness due to factors such as position, the contour length of the area is further calculated. For areas with the same average gray value, the longer the contour, the more likely it is that the area contains particles.

[0020] In an optional embodiment, the foreground score of the image block is obtained by using the average gray value and the contour length of the region, specifically: The average grayscale value is divided by 255 to obtain the normalized grayscale value, and the ratio of the contour length of the area to the perimeter of the area is calculated; The weighted sum of the normalized gray value and the ratio is taken as the foreground score.

[0021] Take the average grayscale value of all pixels in the image block in the particle image. Since the grayscale value is between [0, 255], the normalized grayscale value is obtained by dividing the average grayscale value by 255. The contour length represents the total length of the particle boundary in the image block, and the region perimeter is the boundary perimeter of the image block. The image blocks in feature maps of different scales have different region perimeters in the particle image. The ratio can reduce the influence of the receptive field. By calculating the ratio of the two, the proportion of the particle boundary to the region boundary is evaluated. The foreground score is calculated using a weighted sum method, and the sum of the weights of the normalized grayscale value and the ratio is 1. Since the receptive fields of feature maps of different scales are different, small-scale image blocks represent larger areas in the particle image. In one embodiment, the weight of the ratio of the small-scale feature map to the large-scale feature map is greater. For example, in the large-scale feature map, the weight of the ratio in the foreground score calculation is 0.4, while in the small-scale feature map, the weight of the ratio is 0.5 or 0.6. If there are multiple feature maps of different scales, the weight of the ratio increases with the increase of the scale.

[0022] S2, select multiple target image blocks from each feature map according to the foreground score and the particle size distribution information in the particle image, calculate the attention of the target image block at each layer of the encoder, and then calculate the attention of the target image block and the image block filtered based on the average gray value; The foreground score indicates the possibility of containing particles in the image block. The larger the foreground score, the more likely the image block is to contain particles. When using sparse attention calculation, the image block with a large foreground score will be selected. However, since the number of large-sized particles and small-sized particles in the particle image is not fixed, if the target image block is determined only by the foreground score, it is very likely that some small-sized or large-sized particles will be lost. The present invention further determines the number of target image blocks selected from each feature map based on the particle size distribution in the particle image. In one embodiment, the method of selecting multiple target image blocks from each feature map based on the foreground score and the particle size distribution information in the particle image is specifically as follows: The connected component labeling method is used to calculate the area of ​​each white region in the binary particle image; Cluster the white areas according to their areas. The number of clusters in the cluster is the same as the number of feature maps. Calculate the average area of ​​the white areas in each cluster. Establish the correspondence between the feature maps and clusters based on the average area. The larger the feature map size, the smaller the average area of ​​the corresponding cluster. The particle size distribution information in the particle image is obtained by calculating the ratio of the number of white areas in the cluster corresponding to the feature map; The number of target image blocks of each feature map is obtained according to the particle size distribution information in the particle image and the preset number of target image blocks, and multiple image blocks with the largest foreground scores are selected from the image blocks of each feature map as the target image blocks of the feature map, wherein the number of the multiple image blocks is equal to the number of target image blocks of the feature map.

[0023] After the particle image taken by the electron microscope is binarized, Figure 3 As shown, the connected component labeling method is used to identify each particle (i.e., the white area) in the image, and the area of ​​each particle is calculated, so that the size information of the particle is obtained; in another embodiment, it is not limited to the connected component labeling method and other methods can be used as long as the approximate area of ​​each particle can be roughly identified. Then, the particles are clustered according to the areas of these particles, and the number of clusters is equal to the number of feature maps extracted by the backbone network. Particles with large areas will be divided into one cluster, and particles with small areas will be divided into another cluster. By calculating the average area of ​​particles in each cluster, the corresponding relationship between these clusters and feature maps is established. Specifically, small feature maps correspond to clusters with larger average cluster areas, and feature maps with larger sizes correspond to clusters with smaller average cluster areas. For example, if there are 3 feature maps, the white areas are clustered according to the area. The number of clusters is 3, where the first cluster is a small white area, the second cluster is a medium white area, and the third cluster is a large white area. The first cluster corresponds to the first feature map, the second cluster corresponds to the second feature map, and the third cluster corresponds to the third feature map, where the size of the first feature map is larger than that of the second feature map, and the size of the second feature map is larger than that of the third feature map. Further assuming that there are 4 white areas in the first cluster, 6 white areas in the second cluster, and 8 white areas in the third cluster, the ratio is 4:6:8, that is, the distribution of small particles, medium particles, and large particles in the particle image is approximately 2:3:4. When selecting target image blocks, assuming that a total of 18 target blocks are to be selected, 4 will be selected from the image blocks of the first feature map, 6 will be selected from the image blocks of the second feature map, and 8 will be selected from the image blocks of the third feature map. Specifically, image blocks are selected from each feature map according to the foreground score, and at least one image block with the largest foreground score is selected as the target image block. For example, if there are 4 image blocks to be selected from the first feature map, the 4 image blocks with the largest scores among all the image blocks in the first feature map are selected as the target image blocks. Similarly, the 6 image blocks with the largest foreground scores are selected from the second feature map as the target image blocks.

[0024] In the encoder, each layer of the encoder first calculates the attention of the target image block, and then calculates the attention of the target image block and the image block filtered based on the average gray value. Here, filtering the image block based on the average gray value is mainly to filter out the image blocks that obviously belong to the background, thereby reducing the amount of calculation and the interference to the foreground image block. In an optional embodiment, the calculation of the attention of the target image block and the image block filtered based on the average gray value is specifically as follows; The average gray value of the area that does not contain powder coating particles in the training sample is used as the threshold. If the average gray value of the image block is less than the threshold, the image block is filtered out; The target image patch is used to get the query, and the filtered image patch is used to get the key and value, and the attention is calculated.

[0025] The image blocks are screened by taking the average gray value of the area that does not contain powder coating particles in the training sample as the filtering threshold. Specifically, if the average gray value of an image block is lower than the threshold, it means that the image block may mainly contain the background area, so it is directly filtered out. The retained target image block is processed, and the query (Query) is obtained by using the target image block, and the key (Key) and value (Value) are obtained by using the filtered image block. The attention weight is calculated between the query and the key, and the corresponding value is combined to complete the calculation of the attention mechanism. Then focus on the area containing particle information to avoid the interference of the background area on the model's attention allocation, thereby improving the model's recognition and analysis capabilities for particle images. In one embodiment, the target image block is expanded as a query, or the target image block and the query weight are multiplied and expanded as a query; the filtered image block is expanded as a key and value. It should be noted that in addition to the above two attention calculations, each layer of the encoder will also include other calculations, such as Linear. In addition, after the first attention calculation of each layer, the target image block will be updated, and the updated target image block and the filtered image block will be used to calculate the second attention.

[0026] In one embodiment, the filtered image block is deleted from the image block sequence, and other layers of the encoder do not need to perform the operation of filtering the image block again when performing calculations. Alternatively, the operation of filtering the image block can be placed after the target image block is determined, and the filtered image block and the target image block are input into the encoder, so that each layer of the encoder does not need to perform the filtering operation again.

[0027] In another embodiment, filtering operations are performed at each layer of the encoder, so that as the encoder is executed, more and more image blocks will be deleted and the amount of computation will continue to decrease.

[0028] S3, obtaining the output corresponding to the target image block in the encoder and using it as the input of the decoder, obtaining the particle frame in the particle image through the decoder, and obtaining the particle size information of the powder coating according to the particle frame.

[0029] After being encoded by the encoder, the encoder outputs the target image block, that is, only the output of the target image block sequence number is retained, and the target image block is input into the decoder, and the decoder and the detection head jointly determine the particle frame in the particle image. In order to obtain the particle size information of the powder coating in more detail, in one embodiment, the particle size information of the powder coating is obtained according to the particle frame, specifically: Obtaining a vertical central axis of the particle frame, rotating the central axis according to a preset step length, and counting the length of the central axis in the particle frame after each rotation where the grayscale value is greater than the average grayscale value of the particle frame; After the rotation is completed, the maximum length is taken as the length of the powder coating particle, and the minimum length is taken as the width of the powder coating particle; The length and width of all particles in the particle image are counted to obtain the length distribution histogram and width distribution histogram.

[0030] For each particle frame, first calculate its vertical central axis, and rotate the central axis around the center point of the particle frame in sequence according to a preset angle step, such as 5 degrees or 10 degrees. After each rotation, measure the length of the portion of the central axis inside the particle frame where the grayscale value is higher than the average grayscale value of the particle frame. In an optional embodiment, measure the continuous and maximum length of the portion of the central axis inside the particle frame where the grayscale value is higher than the average grayscale value of the particle frame, that is, the interval must be continuous and maximum to be considered as long or wide. After the rotation is completed, the longest measured length is regarded as the length of the particle, and the shortest measured length is regarded as the width of the particle. This method captures the main extension direction and the minimum extension direction of the particle through multi-angle rotation, and can adapt to the irregularity of the particle shape. Repeat the above steps for all particle frames in the image to obtain the length and width of each particle, summarize the length and width of all particles, draw the length distribution histogram and the width distribution histogram respectively, analyze the distribution law of particle size, and then provide a reference for quality control or material performance evaluation. Figure 4 A length distribution histogram is shown.

[0031] Secondly, the present invention also provides a powder coating particle size automatic analysis system, the system comprising the following modules: A feature extraction module is used to obtain a powder coating particle image taken by an electron microscope, perform feature extraction on the particle image to obtain a plurality of feature maps, divide each feature map into image blocks of the same size, obtain an area of ​​the image block on the particle image, and obtain a foreground score of the image block using an average gray value and a contour line length of the area; An encoding module is used to select multiple target image blocks from each feature map according to the foreground score and the particle size distribution information in the particle image, and first calculate the attention of the target image block at each layer of the encoder, and then calculate the attention of the target image block and the image block filtered based on the average gray value; The decoding and analysis module is used to obtain the output corresponding to the target image block in the encoder and use it as the input of the decoder. The particle frame in the particle image is obtained through the decoder, and the particle size information of the powder coating is obtained according to the particle frame.

[0032] Preferably, the foreground score of the image block is obtained by using the average gray value and the contour length of the region, specifically: The average grayscale value is divided by 255 to obtain the normalized grayscale value, and the ratio of the contour length of the area to the perimeter of the area is calculated; The weighted sum of the normalized gray value and the ratio is taken as the foreground score.

[0033] Preferably, the step of selecting a plurality of target image blocks from each feature image according to the foreground score and the particle size distribution information in the particle image is as follows: The connected component labeling method is used to calculate the area of ​​each white region in the binary particle image; Cluster the white areas according to their areas. The number of clusters in the cluster is the same as the number of feature maps. Calculate the average area of ​​the white areas in each cluster. Establish the correspondence between the feature maps and clusters based on the average area. The larger the feature map size, the smaller the average area of ​​the corresponding cluster. The particle size distribution information in the particle image is obtained by calculating the ratio of the number of white areas in the cluster corresponding to the feature map; The number of target image blocks of each feature map is obtained according to the particle size distribution information in the particle image and the preset number of target image blocks, and multiple image blocks with the largest foreground scores are selected from the image blocks of each feature map as the target image blocks of the feature map, wherein the number of the multiple image blocks is equal to the number of target image blocks of the feature map.

[0034] Preferably, the calculation of the attention of the target image block and the image block filtered based on the average gray value is specifically as follows: The average gray value of the area that does not contain powder coating particles in the training sample is used as the threshold. If the average gray value of the image block is less than the threshold, the image block is filtered out; The target image patch is used to get the query, and the filtered image patch is used to get the key and value, and the attention is calculated.

[0035] Preferably, the particle size information of the powder coating is obtained according to the particle frame, specifically: Obtaining a vertical central axis of the particle frame, rotating the central axis according to a preset step length, and counting the length of the central axis in the particle frame after each rotation where the grayscale value is greater than the average grayscale value of the particle frame; After the rotation is completed, the maximum length is taken as the length of the powder coating particle, and the minimum length is taken as the width of the powder coating particle; The length and width of all particles in the particle image are counted to obtain the length distribution histogram and width distribution histogram.

[0036] Of course, those skilled in the art can understand that all or part of the processes in the above-mentioned embodiments can be implemented by instructing related hardware (such as a processor, a controller, etc.) through a computer program, and the program can be stored in a computer-readable storage medium that can be read by a computer, and the program can include the processes of the above-mentioned method embodiments when executed. The computer-readable storage medium can be a memory, a disk, an optical disk, etc.

[0037] It should be understood that the application of the present invention is not limited to the above examples. For ordinary technicians in this field, improvements or changes can be made based on the above description. All these improvements and changes should fall within the scope of protection of the claims attached to the present invention.

Claims

1. A method for automatically analyzing the particle size of powder coatings, characterized in that: The method comprises the following steps: Obtaining a powder coating particle image photographed by an electron microscope, performing feature extraction on the particle image to obtain a plurality of feature maps, dividing each feature map into image blocks of the same size, obtaining an area of ​​the image block on the particle image, and obtaining a foreground score of the image block using an average gray value and a contour line length of the area; Select multiple target image blocks from each feature map according to the foreground score and the particle size distribution information in the particle image. At each layer of the encoder, the attention of the target image block is first calculated, and then the attention of the target image block and the image block filtered based on the average gray value is calculated. The output corresponding to the target image block in the encoder is obtained and used as the input of the decoder. The particle frame in the particle image is obtained through the decoder, and the particle size information of the powder coating is obtained according to the particle frame.

2. The method according to claim 1, characterized in that The foreground score of the image block is obtained by using the average gray value and the contour length of the region, specifically: The average grayscale value is divided by 255 to obtain the normalized grayscale value, and the ratio of the contour length of the area to the perimeter of the area is calculated; The weighted sum of the normalized gray value and the ratio is taken as the foreground score.

3. The method according to claim 1, characterized in that The method of selecting multiple target image blocks from each feature map according to the foreground score and the particle size distribution information in the particle image is specifically as follows: The connected component labeling method is used to calculate the area of ​​each white region in the binary particle image; Cluster the white areas according to their areas. The number of clusters in the cluster is the same as the number of feature maps. Calculate the average area of ​​the white areas in each cluster. Establish the correspondence between the feature maps and clusters based on the average area. The larger the feature map size, the smaller the average area of ​​the corresponding cluster. The particle size distribution information in the particle image is obtained by calculating the ratio of the number of white areas in the cluster corresponding to the feature map; The number of target image blocks of each feature map is obtained according to the particle size distribution information in the particle image and the preset number of target image blocks, and multiple image blocks with the largest foreground scores are selected from the image blocks of each feature map as the target image blocks of the feature map, wherein the number of the multiple image blocks is equal to the number of target image blocks of the feature map.

4. The method according to claim 1, characterized in that The calculation of the attention of the target image block and the image block after filtering based on the average gray value is specifically as follows: The average gray value of the area that does not contain powder coating particles in the training sample is used as the threshold. If the average gray value of the image block is less than the threshold, the image block is filtered out; The target image patch is used to get the query, and the filtered image patch is used to get the key and value, and the attention is calculated.

5. The method according to claim 1, characterized in that The particle size information of the powder coating is obtained according to the particle frame, specifically: Obtaining a vertical central axis of the particle frame, rotating the central axis according to a preset step length, and counting the length of the central axis in the particle frame after each rotation where the grayscale value is greater than the average grayscale value of the particle frame; After the rotation is completed, the maximum length is taken as the length of the powder coating particle, and the minimum length is taken as the width of the powder coating particle; The length and width of all particles in the particle image are counted to obtain the length distribution histogram and width distribution histogram.

6. An automatic analysis system for powder coating particle size, characterized in that: The system includes the following modules: A feature extraction module is used to obtain a powder coating particle image taken by an electron microscope, perform feature extraction on the particle image to obtain a plurality of feature maps, divide each feature map into image blocks of the same size, obtain an area of ​​the image block on the particle image, and obtain a foreground score of the image block using an average gray value and a contour line length of the area; An encoding module is used to select multiple target image blocks from each feature map according to the foreground score and the particle size distribution information in the particle image, and first calculate the attention of the target image block at each layer of the encoder, and then calculate the attention of the target image block and the image block filtered based on the average gray value; The decoding and analysis module is used to obtain the output corresponding to the target image block in the encoder and use it as the input of the decoder. The particle frame in the particle image is obtained through the decoder, and the particle size information of the powder coating is obtained according to the particle frame.

7. The system according to claim 6, characterized in that The foreground score of the image block is obtained by using the average gray value and the contour length of the region, specifically: The average grayscale value is divided by 255 to obtain the normalized grayscale value, and the ratio of the contour length of the area to the perimeter of the area is calculated; The weighted sum of the normalized gray value and the ratio is taken as the foreground score.

8. The system according to claim 6, characterized in that The method of selecting multiple target image blocks from each feature map according to the foreground score and the particle size distribution information in the particle image is specifically as follows: The connected component labeling method is used to calculate the area of ​​each white region in the binary particle image; Cluster the white areas according to their areas. The number of clusters in the cluster is the same as the number of feature maps. Calculate the average area of ​​the white areas in each cluster. Establish the correspondence between the feature maps and clusters based on the average area. The larger the feature map size, the smaller the average area of ​​the corresponding cluster. The particle size distribution information in the particle image is obtained by calculating the ratio of the number of white areas in the cluster corresponding to the feature map; The number of target image blocks of each feature map is obtained according to the particle size distribution information in the particle image and the preset number of target image blocks, and multiple image blocks with the largest foreground scores are selected from the image blocks of each feature map as the target image blocks of the feature map, wherein the number of the multiple image blocks is equal to the number of target image blocks of the feature map.

9. The system according to claim 6, characterized in that The calculation of the attention of the target image block and the image block after filtering based on the average gray value is specifically as follows: The average gray value of the area that does not contain powder coating particles in the training sample is used as the threshold. If the average gray value of the image block is less than the threshold, the image block is filtered out; The target image patch is used to get the query, and the filtered image patch is used to get the key and value, and the attention is calculated.

10. The system according to claim 6, characterized in that The particle size information of the powder coating is obtained according to the particle frame, specifically: Obtaining a vertical central axis of the particle frame, rotating the central axis according to a preset step length, and counting the length of the central axis in the particle frame after each rotation where the grayscale value is greater than the average grayscale value of the particle frame; After the rotation is completed, the maximum length is taken as the length of the powder coating particle, and the minimum length is taken as the width of the powder coating particle; The length and width of all particles in the particle image are counted to obtain the length distribution histogram and width distribution histogram.

Images