Deformed camouflage spot generation method based on superpixel segmentation and fuzzy C-means clustering

Abstract

The invention discloses a deformed camouflage spot generation method based on superpixel segmentation and fuzzy C-means clustering, and the method specifically comprises the steps: firstly obtaining a background picture, and cutting the background picture into a proportion which is the same as that of a target; then calculating the number of segmented superpixels, performing superpixel segmentation on the background picture, replacing other colors in the superpixels with an average color, clustering the colors of the superpixels, and replacing the colors in the picture with a camouflage main color to obtain a large-spot deformed camouflage pattern; performing edge extraction on the clustered picture to obtain a large-spot deformation camouflage spot pattern, extracting large-spot shapes of the same color, and generating small spots in each large spot by adopting an algorithm combining superpixel segmentation and fuzzy C-means clustering; and finally, superposing the small spot patterns obtained by the large spot patterns to obtain the spot pattern of the small spot deformation camouflage. According to the method, the problems of connection and islanding of the spots are solved, the organic property of combination of the irregular spots is improved, and the design requirement of the deformed camouflage is met.

Description

Technical Field

[0001] The invention relates to the technical field of digital image processing, and in particular to a method for generating deformed camouflage spots based on superpixel segmentation and fuzzy C-means clustering. Background Art

[0002] Deformed camouflage is a multi-color camouflage composed of spots of irregular shapes and sizes. When the target is in different backgrounds, on the one hand, the irregular spots of the deformed camouflage distort the straight outline of the target. On the other hand, the shape and color design of the spots creates a fusion effect with the background or a brightness highlighting effect, making the target produce a camouflaged deformation effect. It is suitable for camouflage protection when mobile targets are active in multiple different backgrounds.

[0003] The irregular shape spots of deformed camouflage are designed according to certain principles, including asymmetry, no right angles, no long straight lines, etc. With the development of computer technology, camouflage design is combined with computer image processing technology. Computer language hopes to obtain a set of accurate data to represent a certain spot. Some researchers use feature vectors to represent irregular spot shapes and establish corresponding spot libraries. Establishing irregular spot libraries can indeed obtain many irregular spots, but the combination of irregular spots has become a problem. For example, two protruding spots cannot be perfectly connected when they touch, and other boundaries outside the contact point will produce blanks. Some researchers use the fractal Brown model to directly generate a whole piece of camouflage pattern to solve the problem of combining spots. However, the Brown fractal model itself is suitable for simulating high-resolution terrain. The flat ground is simulated as a concave and convex ground. The intersection lines obtained by cutting the terrain with contour surfaces will form some irregular spots, but the sizes of these spots are different and cannot fully meet the requirements of spot size. Summary of the invention

[0004] The purpose of the present invention is to provide a method for generating deformed camouflage spots based on superpixel segmentation and fuzzy C-means clustering, so as to solve the connection and island problems of irregular spots, improve the organic nature of the combination between irregular spots, and meet the requirements of deformed camouflage design.

[0005] The technical solution to achieve the purpose of the present invention is: a method for generating deformable camouflage spots based on superpixel segmentation and fuzzy C-means clustering, comprising the following steps:

[0006] Step 1, get a background picture of a natural background;

[0007] Step 2: Crop the background image to the same aspect ratio as the target to be disguised;

[0008] Step 3: According to the area of ​​the target and the reconnaissance resolution of the reconnaissance equipment, the number of superpixel segmentations K is calculated, and the background image is segmented into superpixels;

[0009] Step 4: In each superpixel, replace other colors with the average color to obtain a superpixel pattern of the average color;

[0010] Step 5: Cluster the colors of the superpixels into M colors, replace the colors of the clustered superpixels with the colors of the cluster centers, generate a large spot deformed camouflage pattern with M colors, and replace the M colors in the figure with the main camouflage colors selected as needed to obtain the corresponding large spot deformed camouflage pattern;

[0011] Step 6: extract the edges of the clustered images to obtain a spot pattern of large spot deformed camouflage;

[0012] Step 7: assign a unique label to each color patch. All pixels in the same color patch have the same label. Use the label to extract the large patch shape of the same color to obtain the background image of the large patch shape of each color.

[0013] Step 8: Use an algorithm combining superpixel segmentation and fuzzy C-means clustering to generate small spots in each large spot;

[0014] Step 9: Superimpose the small spot patterns obtained from each large spot to obtain a small spot deformed camouflage pattern.

[0015] Furthermore, the step 1 of obtaining a background image of a natural background is as follows:

[0016] A background picture of a natural background is obtained through an image acquisition device. The picture only contains natural scenery and does not contain artificial targets.

[0017] Furthermore, in step 2, the background image is cropped to have the same aspect ratio as the target to be disguised, as follows:

[0018] Measure the size of the target to be disguised, calculate the aspect ratio of the target, and crop the background image to the corresponding target ratio. The cropped background image contains N pixels.

[0019] Furthermore, in step 3, the number K of superpixel segmentations is calculated according to the area of ​​the target and the reconnaissance resolution of the reconnaissance device, and the background image is segmented into superpixels, as follows:

[0020] The superpixel segmentation SLIC algorithm is used to segment the cropped background image into K superpixels of uniform size. The size of the superpixel S = sqrt(N / K). The spot size of the deformable camouflage is D is the observation distance;

[0021] To ensure that there is no spatial color mixing between the spots of the deformable camouflage, the minimum size d of the deformable camouflage spot is equal to the value of the reconnaissance resolution. The area A of the target to be camouflaged is determined, and the number of superpixel segmentations K = A / d is calculated. 2 ;

[0022] The spot size d of the deformable camouflage is organically combined with the number of superpixel segmentation K. By adjusting K, the designed spot pattern meets the size requirements of the deformable camouflage spots.

[0023] Furthermore, in step 5, the value of the number M of color cluster types is M=3.

[0024] Furthermore, in step 7, a unique label is assigned to each color patch, and all pixels in the same color patch have the same label. The labels are used to extract the large patch shapes of the same color to obtain the background image of the large patch shapes of each color, as follows:

[0025] Step 7.1. For each color patch, create a matrix with the same size as the original image as a mask.

[0026] Step 7.2, find the position of the color patch that needs to be left in the mask matrix, set its value to 1, and set the rest of the positions to 0;

[0027] Step 7.3, multiply the background image by the mask. When the median value of the mask matrix is ​​1, the background image does not change; when the median value of the mask matrix is ​​0, the background image becomes black, thereby extracting the background image of the corresponding shape.

[0028] Furthermore, the algorithm described in step 8 is used to combine superpixel segmentation with fuzzy C-means clustering to generate small spots in each large spot, as follows:

[0029] Step 8.1, according to the ratio of the camouflage unit area of ​​the small spot to the camouflage unit area of ​​the large spot, calculate the number K' of superpixel segmentation of the small spot, and perform superpixel segmentation on the background image;

[0030] Step 8.2: In each superpixel of a small spot, replace other colors with the average color to obtain a superpixel pattern of the average color of the small spot;

[0031] Step 8.3, cluster the colors of the superpixels of small spots into M' colors, replace the color of the superpixel with the color of the cluster center, generate a small spot deformed camouflage pattern with M' colors, and replace the M' colors in the figure with the main camouflage color selected as needed to obtain the corresponding small spot deformed camouflage.

[0032] Furthermore, in step 8.1, the number K' of superpixel segmentation of small spots is 4 times the number K of superpixel segmentation of large spots, that is, K'=4K.

[0033] Furthermore, in step 8.3, the value of the number of color types M' of the superpixel of the small spot is M'=2.

[0034] Furthermore, the small spot patterns obtained from the large spots described in step 9 are superimposed to obtain a small spot deformed camouflage pattern, which is specifically as follows:

[0035] After the small spot pattern of the deformed camouflage is generated, the color in the two-color clustering map is replaced by the main color of the camouflage, and different spots are superimposed to generate a small spot deformed camouflage pattern.

[0036] Compared with the prior art, the present invention has the following significant advantages:

[0037] (1) The spot pattern of the small spot deformable camouflage designed by the present invention combines the K of the superpixel with the size of the spot, and follows the spot size principle of the deformable camouflage;

[0038] (2) A single spot conforms to the principle of irregular spots. The entire camouflage spot pattern is generated by the overall background, which solves the problems of spot connection and island shape, improves the organic nature of the combination between irregular spots, and meets the requirements of deformed camouflage design. BRIEF DESCRIPTION OF THE DRAWINGS

[0039] Figure 1 It is a flow chart of the method for generating deformed camouflage spots based on superpixel segmentation and fuzzy C-means clustering of the present invention.

[0040] Figure 2 It is a grass background image used to generate a camouflage spot pattern in an embodiment of the present invention.

[0041] Figure 3 This is an embodiment of the present invention, in which an algorithm combining superpixel segmentation and fuzzy C-means clustering is used to generate a deformed camouflage image with large spots, wherein (a) is the original image, (b) is a spot boundary image generated by superpixel segmentation, (c) is a three-color clustering image generated by fuzzy C-means, and (d) is a deformed camouflage spot image obtained by performing edge extraction on the clustered image.

[0042] Figure 4 This is a large spot deformed camouflage pattern obtained by replacing the colors in the pattern with the main camouflage color in the embodiment of the present invention.

[0043] Figure 5It is a background image of large spot shapes of each color obtained by extracting large spot shapes of the same color in an embodiment of the present invention, wherein (a) is a grass background image for generating deformed camouflage spots, (b) is a grass background image extracted by a mask corresponding to main color 1, (c) is a grass background image extracted by a mask corresponding to main color 2, and (d) is a grass background image extracted by a mask corresponding to main color 3.

[0044] Figure 6 In the embodiment of the present invention, an algorithm combining superpixel segmentation and fuzzy C-means clustering is used to generate a deformed camouflage image of small spots in large spots of main color 1, wherein (a) is a spot shape background image of main color 1, (b) is a superpixel segmentation image of main color 1, (c) is a two-color clustering image of main color 1, and (d) is a small spot image of main color 1.

[0045] Figure 7 In an embodiment of the present invention, an algorithm combining superpixel segmentation and fuzzy C-means clustering is used to generate a deformed camouflage image of small spots in large spots of main color 2, wherein (a) is a spot shape background image of main color 2, (b) is a superpixel segmentation image of main color 2, (c) is a two-color clustering image of main color 2, and (d) is a small spot image of main color 2.

[0046] Figure 8 In the embodiment of the present invention, an algorithm combining superpixel segmentation and fuzzy C-means clustering is used to generate a deformed camouflage image of small spots in large spots of the main color 3, wherein (a) is a spot shape background image of the main color 3, (b) is a superpixel segmentation image of the main color 3, (c) is a two-color clustering image of the main color 3, and (d) is a small spot image of the main color 3.

[0047] Fig. 9 It is a spot map of small spot deformed camouflage obtained by replacing the colors in the small spot two-color clustering map with the main camouflage color in the embodiment of the present invention, wherein (a) is a large patch camouflage map with main color 1, (b) is a small spot camouflage map with main color 1, (c) is a large patch camouflage map with main color 2, (d) is a small spot camouflage map with main color 2, (e) is a large patch camouflage map with main color 3, and (f) is a small spot camouflage map with main color 3.

[0048] Fig.10 It is a small spot deformed camouflage spot pattern obtained by superimposing small spot patterns obtained from various large spot patterns in the embodiment of the present invention. DETAILED DESCRIPTION

[0049] In view of the design problem of deformable camouflage spot patterns, the present invention combines superpixel segmentation with fuzzy C-means clustering, adopts a method of generating deformable camouflage as a whole, avoids the spot combination problem of generating deformable camouflage from a spot library, and the resulting camouflage pattern meets the spot shape and size requirements of deformable camouflage. The deformable camouflage pattern generated by the present invention has a single spot that conforms to the irregular principle of the pattern, and the camouflage at the edge of the target is composed of spots of different colors, which can distort and change the shape of the target.

[0050] The present invention will be further described in detail below in conjunction with the accompanying drawings and specific embodiments.

[0051] Combination Figure 1 The present invention provides a method for generating deformed camouflage spots based on superpixel segmentation and fuzzy C-means clustering, comprising the following steps:

[0052] Step 1, get a background picture of a natural background;

[0053] Step 2: Crop the background image to the same aspect ratio as the target to be disguised;

[0054] Step 3: According to the area of ​​the target and the reconnaissance resolution of the reconnaissance equipment, the number of superpixel segmentations K is calculated, and the background image is segmented into superpixels;

[0055] Step 4: In each superpixel, replace other colors with the average color to obtain a superpixel pattern of the average color;

[0056] Step 5: Cluster the colors of the superpixels into M colors, replace the colors of the clustered superpixels with the colors of the cluster centers, generate a large spot deformed camouflage pattern with M colors, and replace the M colors in the figure with the main camouflage colors selected as needed to obtain the corresponding large spot deformed camouflage pattern;

[0057] Step 6: extract the edges of the clustered images to obtain a spot pattern of large spot deformed camouflage;

[0058] Step 7: assign a unique label to each color patch. All pixels in the same color patch have the same label. Use the label to extract the large patch shape of the same color to obtain the background image of the large patch shape of each color.

[0059] Step 8: Use an algorithm combining superpixel segmentation and fuzzy C-means clustering to generate small spots in each large spot;

[0060] Step 9: Superimpose the small spot patterns obtained from each large spot to obtain a small spot deformed camouflage pattern.

[0061] As a specific example, the step 1 of obtaining a background image of a natural background is as follows:

[0062] A background picture of a natural background is obtained through an image acquisition device. The picture only contains natural scenery and does not contain artificial targets.

[0063] As a specific example, the background image described in step 2 is cropped to the same aspect ratio as the target to be disguised, as follows:

[0064] Measure the size of the target to be disguised, calculate the aspect ratio of the target, and crop the background image to the corresponding target ratio. The cropped background image contains N pixels.

[0065] As a specific example, in step 3, the number K of superpixel segmentations is calculated according to the area of ​​the target and the reconnaissance resolution of the reconnaissance device, and the background image is segmented into superpixels, as follows:

[0066] The superpixel segmentation SLIC algorithm is used to segment the cropped background image into K superpixels of uniform size. The size of the superpixel S = sqrt(N / K). The spot size of the deformable camouflage is D is the observation distance;

[0067] To ensure that there is no spatial color mixing between the spots of the deformable camouflage, the minimum size d of the deformable camouflage spot is equal to the value of the reconnaissance resolution. The area A of the target to be camouflaged is determined, and the number of superpixel segmentations K = A / d is calculated. 2 ;

[0068] The spot size d of the deformable camouflage is organically combined with the number of superpixel segmentation K. By adjusting K, the designed spot pattern meets the size requirements of the deformable camouflage spots.

[0069] Deformed camouflage needs to ensure that different spots are clearly visible under a certain reconnaissance resolution, so that some spots are integrated into the background, and some spots are distinguished from the background, so as to achieve the effect of changing the original shape of the target as a whole. Therefore, there is a minimum requirement for the size of the spots, which is equal to the value of the reconnaissance resolution. This is to ensure that the deformed camouflage spots are visible under such reconnaissance conditions. If the spot size is too small, the spots will be spatially mixed. For example, if a car is painted with camouflage, if the spots are too small, the original red car may become a yellow car when viewed from a distance (the camouflage mixed color becomes yellow when viewed from a distance). At this time, the deformation effect is lost, and the shape is still the original car shape.

[0070] As a specific example, the value of the number M of color cluster types in step 5 is M=3.

[0071] As a specific example, step 7 assigns a unique label to each color patch, and all pixels in the same color patch have the same label. The labels are used to extract the large patch shapes of the same color to obtain the background image of the large patch shape of each color, as follows:

[0072] Step 7.1. For each color patch, create a matrix with the same size as the original image as a mask.

[0073] Step 7.2, find the position of the color patch that needs to be left in the mask matrix, set its value to 1, and set the rest of the positions to 0;

[0074] Step 7.3, multiply the background image by the mask. When the median value of the mask matrix is ​​1, the background image does not change; when the median value of the mask matrix is ​​0, the background image becomes black, thereby extracting the background image of the corresponding shape.

[0075] As a specific example, the algorithm combining superpixel segmentation and fuzzy C-means clustering described in step 8 is used to generate small spots in each large spot, as follows:

[0076] Step 8.1, according to the ratio of the camouflage unit area of ​​the small spot to the camouflage unit area of ​​the large spot, calculate the number K' of superpixel segmentation of the small spot, and perform superpixel segmentation on the background image;

[0077] Step 8.2: In each superpixel of a small spot, replace other colors with the average color to obtain a superpixel pattern of the average color of the small spot;

[0078] Step 8.3, cluster the colors of the superpixels of small spots into M' colors, replace the color of the superpixel with the color of the cluster center, generate a small spot deformed camouflage pattern with M' colors, and replace the M' colors in the figure with the main camouflage color selected as needed to obtain the corresponding small spot deformed camouflage.

[0079] As a specific example, the number K' of superpixel segmentation of small spots in step 8.1 is 4 times the number K of superpixel segmentation of large spots, that is, K'=4K.

[0080] As a specific example, the value of the number of color types M' of the superpixel of the small spot in step 8.3 is M'=2.

[0081] As a specific example, the small spot patterns obtained from the large spots described in step 9 are superimposed to obtain a small spot deformed camouflage pattern, which is as follows:

[0082] After the small spot pattern of the deformed camouflage is generated, the color in the two-color clustering map is replaced by the main color of the camouflage, and different spots are superimposed to generate a small spot deformed camouflage pattern.

[0083] The present invention is further described in detail below with reference to the accompanying drawings and specific embodiments.

[0084] Example

[0085] This embodiment takes a large-scale equipment performing parking, maneuvering, and action tasks in a typical eastern background as the application scenario, and designs a deformable camouflage pattern that satisfies the mobile equipment's activities in the area based on the woods, grasslands, farmlands, villages, dirt roads, cement roads, asphalt roads, and other background features that the equipment passes through during its movement.

[0086] This embodiment selects a grass background to generate a camouflage spot pattern, such as Figure 2 As shown, it is a lawn in a park in Laoshan District, Nanjing City, which is acquired by a drone at a height of 5 meters, with a pixel size of 3956*5280. In this embodiment, the grass acquired at a height of 5 meters is selected as the background for generating the camouflage spot pattern, in order to reduce the hardware requirements of this method. Even if the designer does not have a drone, a ladder and a camera can be used instead.

[0087] Due to the sensitivity of a large device, for example, a small car on the market is 5 meters long and 2 meters wide. The resolution of satellite reconnaissance is 0.1 meters. The background image needs to be segmented into 1000 superpixels, that is, K = S / d 2 =(2*5) / (0.1*0.1)=1000. At the same time, the aspect ratio of the background used to generate the camouflage pattern should be consistent with the target. If not, the background image needs to be cropped according to the target ratio. The purpose of cropping is to ensure that the camouflage texture generated later is consistent with the original background without compression and stretching. To ensure that the generated camouflage spots can fit the target, they need to be cropped to a size of 2000*5000, the same as the aspect ratio of a car.

[0088] The cropped image is processed using superpixel segmentation and fuzzy C-means clustering algorithms to generate the corresponding camouflage pattern, such as Figure 3 .

[0089] Figure 3 The algorithm combining superpixel segmentation and fuzzy C-means clustering is used in the embodiment of the present invention to generate a large spot deformed camouflage pattern, wherein Figure 3 (a) in the figure is the original image. Figure 3 (b) in the figure is the spot boundary map generated by superpixel segmentation. Figure 3 (c) in the figure is a three-color clustering graph generated by fuzzy C-means. Figure 3 (d) in the figure is a deformed camouflage spot pattern obtained by performing edge extraction on the clustered image.

[0090] Select the main camouflage color as needed Figure 3 The present invention replaces the three colors from dark to light in the figure with earth color (RGB value: 230, 192, 161), green (124, 128, 64), and dark green (68, 77, 54) to obtain the corresponding large spot deformed camouflage pattern, such as Figure 4 .

[0091] from Figure 4 It can be seen that each spot is an irregular spot when viewed alone, which is consistent with the irregularity of the deformed camouflage; the minimum size of the camouflage spots designed by this method can also meet the visibility of the spots, and no spatial color mixing will occur at the corresponding reconnaissance distance; any edge of the camouflage basically contains 3 different colors, which meets the edge fragmentation. In theory, the deformed camouflage designed by this method should have a good deformation camouflage effect.

[0092] The present invention intends to design a small spot camouflage, which requires Figure 4 For re-segmentation, we first use a mask to extract the large monochrome spots of the deformed camouflage, and then use an algorithm that combines superpixel segmentation with fuzzy C-means clustering to generate small spots in each large spot.

[0093] After completing the design of the large spots of deformed camouflage, each color patch will be assigned a unique label, and the labels of all pixels in the same color spot are the same, so the labels can be used to extract the shape of the large spots of the same color.

[0094] Figure 5 is a background pattern of a large spot shape of each color obtained by extracting the large spot shape of the same color in the embodiment of the present invention, Figure 5 (a) is the grass background pattern for generating deformed camouflage spots. Figure 5 (b) in the figure is the grass background pattern extracted by the mask corresponding to the primary color 1. Figure 5 (c) in the figure is the grass background pattern extracted by the mask corresponding to the primary color 2. Figure 5 (d) in the figure is the grass background pattern extracted by the mask corresponding to the main color 3.

[0095] The size of small spots needs to be smaller than that of large spots, so the size of the camouflage units (superpixels) that make up the small spots must also be smaller than the camouflage units of the large spots. In the present invention, the area of ​​the camouflage units of small spots is 1 / 4 of the camouflage units of large spots. When generating large spot deformed camouflage, the K value of superpixel clustering is 1000, so when generating small spot camouflage, the K value of superpixel clustering is 4000.

[0096] Figure 6 (a) in the figure is the grass background with the main color 1 shape extracted by the mask. Figure 6 (b) in Figure 6 The superpixel segmentation map of (a) in Figure 6 (c) in Figure 6 (b) The two-color image generated by fuzzy clustering does not contain the black color in the image. Figure 6 (d) in Figure 6 The speckle pattern in (c).

[0097] Figure 7 (a) in the figure is the grass background of the main color 2 shape extracted by the mask. Figure 7 (b) in Figure 7 The superpixel segmentation map of (a) in Figure 7 (c) in Figure 7 (b) The two-color image generated by fuzzy clustering does not contain the black color in the image. Figure 7 (d) in Figure 7 The speckle pattern in (c).

[0098] Figure 8 (a) is the grass background of the main color 3 shape extracted by the mask. Figure 8 (b) in Figure 8 The superpixel segmentation map of (a) in Figure 8 (c) in Figure 8 (b) The two-color image generated by fuzzy clustering does not contain the black color in the image. Figure 8 (d) in Figure 8 The speckle pattern in (c).

[0099] After the small spot pattern of the deformed camouflage is generated, you only need to replace the color in the two-color clustering map with the main color of the camouflage, and then superimpose different spots to generate the small spot deformed camouflage.

[0100] When determining the main color of the deformed camouflage, only three colors are selected as the main colors of the deformed camouflage. After each large spot generates a corresponding small spot, two colors are needed to fill it. The small spots after filling need to produce a better spatial color mixing effect when observed from a distance. After conducting many experiments on spatial color mixing, it is concluded that spatial color mixing is related to the colors involved in the color mixing. The smaller the color difference between the two colors, the stronger the color mixing ability. In this embodiment, the original color of the large spot is retained as a color with a larger area among the small spots, and then a color with the closest brightness and color difference to the original color of the large spot is selected as another color of the small spot according to the needs.

[0101] Fig. 9 (a), (c), and (e) are large patches of a single color. Fig. 9 (b), (d), and (f) are the corresponding small spot camouflage patterns. As can be seen from the figure, only Fig. 9 In (b), the large spots of the main color 1 are more obvious than the small spots. Fig. 9 The small spots in the two large spots in (d) and (f) have shown a good spatial color mixing effect.

[0102] Fig.10 This is the effect of three large spots camouflage converted into small spots camouflage, which is an overall small spots camouflage pattern. The two colors in the earth-colored spots are more obvious, and the different colors in the green spots are difficult to distinguish. The effect of spatial color mixing has a lot to do with the colors involved in the mixing. In the case of long distance, the earth-colored spots will also have spatial color mixing. Since the two colors in the earth-colored spots have the same brightness, even if the spatial color mixing effect is not good at a long distance, the earth-colored spots and other spots meet the requirements of the main colors of the deformed camouflage, and will not affect the deformation effect of the camouflage.

[0103] The colors used in the present invention are the main colors of a typical background in a certain eastern region, and users can replace them with colors they need as needed.

[0104] The above are only preferred embodiments of the present invention. It should be pointed out that, for ordinary technicians in this technical field, several improvements and modifications can be made without departing from the principles of the present invention. These improvements and modifications should also be regarded as the scope of protection of the present invention.

Claims

1. A method for generating deformable camouflage spots based on superpixel segmentation and fuzzy C-means clustering, characterized in that: The following steps are involved: Step 1, get a background picture of a natural background; Step 2: Crop the background image to the same aspect ratio as the target to be disguised; Step 3: According to the area of ​​the target and the reconnaissance resolution of the reconnaissance equipment, the number of superpixel segmentations K is calculated, and the background image is segmented into superpixels; Step 4: In each superpixel, replace other colors with the average color to obtain a superpixel pattern of the average color; Step 5: Cluster the colors of the superpixels into M colors, replace the colors of the clustered superpixels with the colors of the cluster centers, generate a large spot deformed camouflage pattern with M colors, and replace the M colors in the figure with the main camouflage colors selected as needed to obtain the corresponding large spot deformed camouflage pattern; Step 6: extract the edges of the clustered images to obtain a spot pattern of large spot deformed camouflage; Step 7: assign a unique label to each color patch. All pixels in the same color patch have the same label. Use the label to extract the large patch shape of the same color to obtain the background image of the large patch shape of each color. Step 8: Use an algorithm combining superpixel segmentation and fuzzy C-means clustering to generate small spots in each large spot; Step 9: Superimpose the small spot patterns obtained from each large spot to obtain a small spot deformed camouflage pattern.

2. The method for generating deformable camouflage spots based on superpixel segmentation and fuzzy C-means clustering according to claim 1, characterized in that: Step 1 is to obtain a background image of a natural background, as follows: A background picture of a natural background is obtained through an image acquisition device. The picture only contains natural scenery and does not contain artificial targets.

3. The method for generating deformable camouflage spots based on superpixel segmentation and fuzzy C-means clustering according to claim 2, characterized in that: In step 2, the background image is cropped to the same aspect ratio as the target to be disguised, as follows: Measure the size of the target to be disguised, calculate the aspect ratio of the target, and crop the background image to the corresponding target ratio. The cropped background image contains N pixels.

4. The method for generating deformable camouflage spots based on superpixel segmentation and fuzzy C-means clustering according to claim 3 is characterized in that: According to step 3, the number of superpixel segmentations K is calculated based on the area of ​​the target and the reconnaissance resolution of the reconnaissance equipment, and the background image is segmented into superpixels, as follows: The superpixel segmentation SLIC algorithm is used to segment the cropped background image into K superpixels of uniform size. The size of the superpixel S = sqrt(N / K). The spot size of the deformable camouflage is D is the observation distance; To ensure that there is no spatial color mixing between the spots of the deformable camouflage, the minimum size d of the deformable camouflage spot is equal to the value of the reconnaissance resolution. The area A of the target to be camouflaged is determined, and the number of superpixel segmentations K = A / d is calculated. 2 ; The spot size d of the deformable camouflage is organically combined with the number of superpixel segmentation K. By adjusting K, the designed spot pattern meets the size requirements of the deformable camouflage spots.

5. The method for generating deformable camouflage spots based on superpixel segmentation and fuzzy C-means clustering according to claim 1, characterized in that: In step 5, the value of the number M of color cluster types is M=3.

6. The method for generating deformable camouflage spots based on superpixel segmentation and fuzzy C-means clustering according to claim 1, characterized in that: As described in step 7, a unique label is assigned to each color patch. All pixels in the same color patch have the same label. The labels are used to extract the large patch shapes of the same color to obtain the background image of the large patch shape of each color, as follows: Step 7.

1. For each color patch, create a matrix with the same size as the original image as a mask. Step 7.2, find the position of the color patch that needs to be left in the mask matrix, set its value to 1, and set the rest of the positions to 0; Step 7.3, multiply the background image by the mask. When the median value of the mask matrix is ​​1, the background image does not change; when the median value of the mask matrix is ​​0, the background image becomes black, thereby extracting the background image of the corresponding shape.

7. The method for generating deformable camouflage spots based on superpixel segmentation and fuzzy C-means clustering according to claim 1, characterized in that: The algorithm described in step 8 that combines superpixel segmentation with fuzzy C-means clustering is used to generate small spots in each large spot, as follows: Step 8.1, according to the ratio of the camouflage unit area of ​​the small spot to the camouflage unit area of ​​the large spot, calculate the number K' of superpixel segmentation of the small spot, and perform superpixel segmentation on the background image; Step 8.2: In each superpixel of a small spot, replace other colors with the average color to obtain a superpixel pattern of the average color of the small spot; Step 8.3, cluster the colors of the superpixels of small spots into M' colors, replace the color of the superpixel with the color of the cluster center, generate a small spot deformed camouflage pattern with M' colors, and replace the M' colors in the figure with the main camouflage color selected as needed to obtain the corresponding small spot deformed camouflage.

8. The method for generating deformable camouflage spots based on superpixel segmentation and fuzzy C-means clustering according to claim 7, characterized in that: In step 8.1, the number K' of superpixel segmentation of small spots is 4 times the number K of superpixel segmentation of large spots, that is, K'=4K.

9. The method for generating deformable camouflage spots based on superpixel segmentation and fuzzy C-means clustering according to claim 7, characterized in that: The value of the number of color types M' of the superpixel of the small spot in step 8.3 is M'=2.

10. The method for generating deformable camouflage spots based on superpixel segmentation and fuzzy C-means clustering according to claim 1, characterized in that: The small spot patterns obtained from each large spot described in step 9 are superimposed to obtain a small spot deformed camouflage pattern, which is as follows: After the small spot pattern of the deformed camouflage is generated, the color in the two-color clustering map is replaced by the main color of the camouflage, and different spots are superimposed to generate a small spot deformed camouflage pattern.

Images