CN112950733B - Multi-terrain camouflage pattern generation method suitable for ground equipment - Google Patents
Multi-terrain camouflage pattern generation method suitable for ground equipment Download PDFInfo
- Publication number
- CN112950733B CN112950733B CN202110225866.2A CN202110225866A CN112950733B CN 112950733 B CN112950733 B CN 112950733B CN 202110225866 A CN202110225866 A CN 202110225866A CN 112950733 B CN112950733 B CN 112950733B
- Authority
- CN
- China
- Prior art keywords
- color
- camouflage
- background
- terrain
- pattern
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Active
Links
Images
Classifications
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T11/00—2D [Two Dimensional] image generation
- G06T11/001—Texturing; Colouring; Generation of texture or colour
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T7/00—Image analysis
- G06T7/10—Segmentation; Edge detection
- G06T7/13—Edge detection
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T7/00—Image analysis
- G06T7/10—Segmentation; Edge detection
- G06T7/136—Segmentation; Edge detection involving thresholding
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T7/00—Image analysis
- G06T7/60—Analysis of geometric attributes
- G06T7/62—Analysis of geometric attributes of area, perimeter, diameter or volume
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T2207/00—Indexing scheme for image analysis or image enhancement
- G06T2207/10—Image acquisition modality
- G06T2207/10024—Color image
Landscapes
- Engineering & Computer Science (AREA)
- Physics & Mathematics (AREA)
- General Physics & Mathematics (AREA)
- Theoretical Computer Science (AREA)
- Computer Vision & Pattern Recognition (AREA)
- Geometry (AREA)
- Image Analysis (AREA)
- Image Processing (AREA)
Abstract
The invention discloses a multi-terrain camouflage pattern generation method suitable for ground equipment, which comprises the following steps: s1, dominant color extraction: selecting a typical terrain background image of a region related to the equipment to be camouflaged, and extracting and calculating the multi-terrain camouflage color and the corresponding area proportion thereof; s2, designing a plaque library: based on the bionics principle, acquiring a background typical biological shape image, and constructing a camouflage patch library through matrixing processing; s3, camouflage pattern generation: and (3) randomly filling the multi-terrain camouflage color and the area ratio thereof in the step (S1) to generate a main color pattern, selecting a biological information patch from the camouflage patch library in the step (S2) to blend the main color pattern to generate a deformed camouflage color, and performing digital processing to obtain the digital camouflage color. The method is simple and quick in calculation, the generated camouflage pattern texture is high in randomness, the size of the camouflage spots is controllable, and the method is suitable for multiple terrain backgrounds. The method and the technical thought of the invention promote the development of the camouflage theory.
Description
Technical Field
The invention belongs to the technical field of image processing and analysis and camouflage design, and particularly relates to a multi-terrain camouflage pattern generation method suitable for ground equipment.
Background
In modern wars, camouflage stealth technology has become an important component of modern battlefields as a high-technology anti-reconnaissance means. The camouflage stealth technology aims to conceal oneself, deceive and confuse enemies, change the detectable information characteristics of objects such as weaponry and equipment and enable an enemy detection system to be difficult to find or to shorten the finding distance. The camouflage painting is used as the simplest camouflage technology, is one of the simplest and most effective military camouflage protection means, and is an important technology for guaranteeing military strength, resisting modern reconnaissance detection and accurately guiding weapon capture.
The existing camouflage pattern design is mostly focused on the research on a single terrain background, the design method is mature, and a good camouflage effect can be achieved, but the camouflage pattern designed by the method is difficult to simultaneously resist the high-precision reconnaissance and fine capture threats of a plurality of terrain backgrounds in the maneuvering process of equipment, is not suitable for camouflage of ground equipment, and for the camouflage pattern design of the ground equipment, the visual characteristic information of the equipment possibly related to regional background images needs to be comprehensively considered, so that the camouflage protection requirement in an information-based battlefield is difficult to meet.
The multi-terrain camouflage color camouflage pattern design is similar to a traditional camouflage color pattern and comprises three parts, namely camouflage color dominant color extraction, patch design and pattern generation. For the extraction of the main color of the camouflage color, peng Bian et al [1] extracts the main color of the background by firstly converting the color space into the HSV color space and then differentiating the color histogram method. Zhang Yong, qi Jia et al [2-3] extract background dominant colors through a K-means clustering method and an improved method thereof, and the study of the scholars on the extraction of single-background camouflage dominant colors has a good effect, but does not provide a solution for extracting multi-background camouflage dominant colors. In the aspect of the problem, weiyu Xuan et al [4] performs clustering algorithm after splicing background images to extract multi-background camouflage main colors, but the main colors extracted by the method may fall into local minimum values. ZHU et al [5] use the frequency distribution and additional weights to perform multiple updates and select the optimal parameters to extract dominant colors, but the method is only applicable to similar terrain. In the plaque design part, hejrandoost et al [6] fuses multi-background plaque images through thinning, xue Feng et al [7] generates camouflage plaques by establishing a template library and arranging the plaques in the library by a greedy algorithm, and the plaque design methods all need to extract and fuse the plaques of each background image and are high in calculation complexity. For pattern generation, a generation method such as random fill, correspondence fill, and fill in which the color of the pixel having the largest number of basic cells is used is generally employed.
Reference to the literature
[1]PENG Bian,JIN Yi,ZHANG Nairen.Fuzzy c-means clustering based digital camouflage pattern design and its evaluation[C]//IEEE International Conference on Signal Processing,Yantai China,2010:1017-1020.
[2] Zhang Yong, wu Wen Jian, liu Shiming, camouflage color selection based on improved K-means clustering analysis [ J ]. Computer engineering and applications, 2009,45 (006): 210-212.
[3]QI Jia,LIN Ziqiang,Hu Jianghua,,et al.Design and Evaluation of Facial Camouflage Pattern[C]//.International Conference on Artificial Intelligence and Computer Science(AICS 2019),Hubei Zhongke Institute of Geology and Environment Technology,2019:4.
[4] Method for extracting theme colors and generating camouflage colors from multiple pictures under the background of visible light: china, CN201910898945.2[ P ].2020-01-21.
[5]ZHU W,ZHANG Y,ZHU H,et al.All-terrain camouflage design on frequency analysis[C]//Control&Decision Conference.IEEE,Yinchuan China,2016:3102-3105.
[6]Hejrandoost AA,Safabakhsh R.Thinning Based Multipurpose Camouflage Pattern Design[C]//Machine Vision&Image Processing.IEEE,Iranian,2012:1-5.
[7]Xue F,Xu S,Luo Y T,et al.Design of digital camouflage by recursive overlapping of pattern templates[J].Neurocomputing,2015,172(JAN.8):262-270.
Disclosure of Invention
The invention aims to solve the problem that the existing camouflage pattern design method of the ground equipment under a single background is difficult to meet the camouflage requirement under the multi-terrain background environment, and provides a multi-terrain deformation and digital camouflage pattern generation method suitable for the ground equipment. The method comprises three modules, namely, dominant color extraction: selecting a typical background image of a to-be-disguised device possibly related to a region, and extracting multi-terrain camouflage color dominant colors and area ratios thereof through differential analysis of background dominant colors; designing a plaque library: based on the bionics principle, typical biological shape pictures are collected, and a camouflage patch library is constructed; and (3) generation of camouflage patterns: and randomly filling and generating a main color pattern based on the multi-terrain camouflage color main color and the area ratio thereof, selecting a biological information patch from the patch library, blending the biological information patch into the camouflage color main color pattern to generate a deformed camouflage color, and performing digital processing to obtain the digital camouflage color.
The purpose of the invention is realized by the following steps:
a method of generating a multi-terrain camouflage pattern for use with surface equipment, comprising the steps of:
s1, extracting a dominant color;
s1.1, analyzing the terrain environment where the ground equipment to be disguised is located, collecting a plurality of terrain background images, and respectively carrying out preprocessing operation on each background image, wherein the preprocessing operation comprises the following steps: size normalization and image denoising; selecting a typical image from each terrain background, and storing the typical image in a sequence P, wherein P = { I = { (I) } i I =1,2,. Cndot.n }, where N denotes the number of selected background images;
s1.2, adopting an image color quantization algorithm to carry out quantization on I in the selected background image sequence i And respectively carrying out color quantization processing, quantizing the color data to K background dominant colors, recording the color values and the corresponding area proportions of the background dominant colors, and recording as follows:
the color value representing the dominant color of the background is
The three-dimensional vector of the composition,
representing an area ratio of the background dominant color;
s1.3. For different background images I i Background dominant color of (1)
Performing color difference analysis, and calculating the main color of the multi-terrain camouflage color and the corresponding area proportion thereof according to the analysis result;
the average color difference formula is used as a measurement function, the color difference between the background dominant colors is analyzed, and the smaller the measurement function value is, the smaller the difference between the two background dominant colors is;
setting and selecting two different topographic background images I i 、I j After the color quantization processing of step S1.2, the color value and the corresponding area ratio of the background dominant color are respectively
The metric function formula can be derived as follows:
finding the background dominant color combination that minimizes the overall color diversity according to the definition of the color similarity metric function,
the optimization process formula is described as follows:
the above formula represents substituting into the background image I j All permutation combinations of the background dominant colors are searched for variables k1', k2', k3',. K, k' which minimize the overall calculation result, namely I is carried out i 1,2,3, K background dominant colors and I j K1, k2, k3,. K background dominant color combinations;
wherein K1', K2', K3', K' is in the scope of {1,2, 3.., K }, and K represents the number of multi-background camouflage dominant colors;
according to the above optimization process, the background dominant color combination with the minimum overall color difference is:
calculating the average value of the above background dominant color combinations to obtain the multi-terrain camouflage dominant color s And area ratio p s S ∈ {1,2,.., K }, the formula is expressed as follows:
s2, designing a plaque library;
based on the principle of bionics, a typical biological shape image of a background is collected, preprocessed and trimmed, and then the biological shape image is converted into a biological information patch by using a matrixing method;
the algorithm is described as follows:
step 1: inputting a biological shape image A m M is equal to {1,2, 3.,. M }, M is the total number of the biological shape images to be processed, the human-computer interactive frame selects the minimum bounding rectangle containing all the biological shapes, and the rectangle is output and named as A _ S m ；
Step 2: for A _ S m Performing size normalization processing to set the size of the size toL, the size is the size of the biological information plaque, then the decolorizing treatment is carried out, and the RGB color image is converted into a gray image;
and step 3: performing image self-adaptive threshold segmentation algorithm on the gray level image to segment out the biological shape contour, and recording the segmentation result as A _ cut m Convert A _ cut m Is a binary matrix, namely the biological information plaque, and is marked as A _ result m A _ result is added m Sequentially storing the data into a plaque library A _ CLUB;
and 4, step 4: circularly executing the steps 1-3 until all the M images containing the biological shape information are processed and stored in the plaque library, and finishing the circular process;
s3, generating camouflage patterns;
s3.1, generating a dominant color pattern;
firstly, dividing the size of the main color pattern and the size of the mosaic unit to obtain the number of the mosaic units in the main color pattern, wherein the formula is as follows:
wherein o x q is the size of the dominant color pattern, and λ x λ is the size of the mosaic unit;
then the number n of mosaic units of each multi-terrain camouflage color dominant color is distributed according to the area proportion calculated in the step S1.3 r ，r∈{1,2,...,K}；
Finally, randomly selecting n in the dominant color pattern r A mosaic unit filled with corresponding multi-terrain camouflage main color s Until all the mosaic units are filled, the main color pattern X is generated;
s3.2, generating a deformed camouflage pattern;
after the dominant color pattern is generated, a biological information patch A _ result is randomly selected from the patch library m The patch features are fused into the dominant color patterns by a mathematical morphology method, and a specific fusion process is carried out by adopting an open-close operation formula, wherein the formula is expressed as follows:
I_com＝X·(XοA_result m )
i _ com represents the generated deformed camouflage pattern, operator. Representing a morphological open operation, an operator, representing a morphological closed operation;
s3.3, generating a digital camouflage pattern;
generating a digital camouflage pattern through digital processing after the deformed camouflage pattern is generated;
firstly, determining the size d of a basic unit of the digital camouflage color, then, processing the deformed camouflage color pattern in blocks, wherein the size of each block is the size of the basic unit of the digital camouflage color, and the processing process of the blocks is as follows:
counting the pixel with the most occurrence times in the block and assigning the color value of the pixel to all the pixels in the block;
the operation is circulated until all blocks in the deformed camouflage pattern are completely executed, and the processed image is the digital camouflage pattern;
further: the method for calculating the dominant color and the area ratio of the multi-background camouflage color in the step S1.3 is not limited to two different terrain backgrounds, and is also suitable for extracting the dominant color of more than two terrain backgrounds. .
Further: step S3.2 can flexibly control the size of the patch in the deformed camouflage pattern by controlling the ratio of the size L of the biological information patch to the size lambda of the mosaic unit in the fusion process, and the size of the mosaic unit is generally set to be smaller than that of the biological information patch in consideration of the fact that the size of the patch in the deformed camouflage pattern cannot be too large, and the size setting can refer to: 0.5 x L is not more than lambda and is less than L.
And further: step S3.3 in the process of blocking, the pixels with the largest number of occurrences in the statistical block follow the following rules:
(1) If only one pixel exists in the block with the largest occurrence frequency, replacing all pixels in the block with the pixel with the largest occurrence frequency;
(2) If the pixels with the most occurrence times in the block have G types, and G belongs to {2, 3., K }, one of the G types of pixels is selected randomly to replace all the pixels in the block.
Further: step S3.3 the blocking process employs sequential operations, i.e. from left to right, top to bottom in the image matrix.
Compared with the prior art, the invention has the beneficial effects that:
(1) The method measures the color similarity among the background dominant colors of the plurality of terrain backgrounds by adopting an average color difference formula, obtains the multi-terrain camouflage color dominant colors and the area proportion thereof by adopting a mean value calculation method according to a measurement result, and fuses the color information of the plurality of terrain backgrounds, so that the dynamic equipment to be disguised has better color fusion degree for each type of terrain, and the dynamic disguising effect is realized.
(2) Based on the bionics principle, the biological shape is used as the digital and deformed camouflage pattern patch, the processing of image patch segmentation, extraction, fusion and the like on each terrain background image is avoided, the calculation simplicity is improved, only direct calling is needed in the subsequent camouflage color generation process of one-time patch library manufacturing, and the real-time performance of camouflage pattern generation is enhanced.
(3) Before the digital codes are generated and the camouflage is deformed, the multi-terrain camouflage main color and the area ratio thereof are used for randomly generating the main color patterns, so that the randomness of the camouflage texture is enhanced, and the camouflage effect is improved.
(4) The method adopts mathematical morphology operation to fuse the biological information patch into the dominant color pattern to generate the deformed camouflage, so that the size of the camouflage patch is flexible and controllable, the mathematical morphology only relates to logical relation operation, the arithmetic operation speed is high, and the camouflage generation efficiency is improved.
(5) According to the method for generating the multi-terrain deformation and digital camouflage patterns, the generated camouflage patterns are suitable for the multi-variability terrain background, reference can be provided for camouflage design of various ground equipment, and the method is also suitable for camouflage pattern design of other military targets under various different terrain backgrounds.
Drawings
FIG. 1 is a block flow diagram of a method for generating a multi-terrain camouflage pattern for use with a piece of land equipment in accordance with the present invention;
FIG. 2 is a diagram of a typical terrain background picture taken in an embodiment of the present invention;
FIG. 3 is a result of the background dominant color after the background image color quantization according to the embodiment of the present invention;
FIG. 4 illustrates multi-terrain camouflage mass tone in accordance with an embodiment of the present invention;
FIG. 5 is a flow chart of a plaque library design algorithm in an embodiment of the present invention;
FIG. 6 is a patch diagram of segmentation results and biological information in the design of a patch library according to an embodiment of the present invention;
FIG. 7 is a dominant color pattern in an embodiment of the present invention;
FIG. 8 is a camouflage pattern of willow leaf, tiger stripe, maple leaf patch in accordance with an embodiment of the present invention;
FIG. 9 is a result of digitizing a deformed camouflage pattern according to an embodiment of the present invention;
Detailed Description
The present invention will be described in further detail with reference to the accompanying drawings and specific embodiments.
Referring to fig. 1, a method for generating a multi-terrain-deformation digital camouflage color suitable for ground equipment includes the following steps:
s1, extracting a dominant color;
s1.1, analyzing the terrain environment where the ground equipment to be disguised is located, collecting a plurality of terrain background images, and respectively carrying out preprocessing operation on the background images, wherein the preprocessing operation comprises the following steps: size normalization and image denoising; selecting a typical image from each terrain background, storing the typical image in a sequence P, wherein P = { I = { (I) } i I =1,2,. N }, N denoting the number of selected background images;
the ground equipment to be disguised mainly comprises ground equipment such as armored vehicles, tanks, military communication vehicles, military transport vehicles and the like, and comprises dynamic and static ground equipment;
the background image preprocessing aims at unifying the size of a background image and removing noise interference of imaging equipment and an external environment on the background image;
typical terrain background images selected in embodiments of the invention are shown in figures 2 (a), (b), forest land and grassland type background respectively, namely the sequence P = { I = { (I) 1 ，I 2 }，N＝2；
Setting a background image size o q =300 x 300, and denoising an image sequence by adopting a wavelet threshold denoising mode;
s1.2, image color quantization algorithm is adoptedFor I in background image sequence i And respectively carrying out color quantization processing, quantizing the color data to K background dominant colors, recording the color values and the corresponding area proportions of the background dominant colors, and recording as follows:
the color value representing the dominant color of the background is
The three-dimensional vector of the composition,
representing an area ratio of the background dominant color;
the background dominant color number K generally ranges from: k is more than or equal to 3 and less than or equal to 6;
based on the visual characteristics of human eyes and the design specifications of the camouflage color, the color of the background picture can be represented by a plurality of dominant colors, and the camouflage effect of the camouflage color is not reduced by using the dominant colors;
the color types of the background images are thousands of, the real-time performance of the generation of the camouflage pattern is considered, and the color types are compressed by adopting an image color quantization algorithm;
the image color quantization algorithm comprises: quantizing a color histogram, K-means clustering, FCM clustering, ISODATA clustering and the like;
considering the real-time and accuracy requirements of the generation of the camouflage pattern, the K-means clustering algorithm which is mature and simple to calculate is preferentially selected in the embodiment, and the initial clustering center selects a point with large chromatic aberration in the image, so that the color quantification accuracy is improved;
in the embodiment of the present invention, the color of the background image is quantized by a K-means clustering algorithm, K =3 is taken, the result of the dominant color of the background is shown in fig. 3 (a) and (b), and the specific color values and area of the dominant color of the background account for the following table 1:
TABLE 1 background dominant color values and area ratios
S1.3. For different background images I i Background dominant color of
Performing color difference analysis, and calculating the main color of the multi-terrain camouflage color and the corresponding area proportion thereof according to the analysis result;
analyzing the color difference between the dominant colors of the backgrounds by adopting an average color difference formula as a measurement function, wherein the smaller the measurement function value is, the smaller the difference between the dominant colors of the two backgrounds is;
setting and selecting two different terrain background images I i 、I j After the color quantization processing of step S1.2, the color value and the corresponding area ratio of the background dominant color are respectively
The metric function equation can be derived as follows:
when the terrain background is N pieces, the color value and the corresponding area proportion of the dominant color of the background are respectively
The metric function is formulated as:
finding the background dominant color combination which minimizes the overall color difference according to the definition of the color similarity metric function,
the optimization process formula is described as follows:
the above formula represents substituting into the background image I j All permutation combinations of the background dominant colors find the variables k1', k2', k3',. K ', k1, k2, k3', k, i.e. perform I, which minimizes the overall computation i 1,2,3, K background dominant colors and I j K1, k2, k3,. K background dominant color combinations;
wherein K1', K2', K3', K' is in the scope of {1,2, 3.., K }, and K represents the number of multi-background camouflage dominant colors;
according to the above optimization process, the background dominant color combination with the minimum overall color difference is:
in the embodiment of the present invention, the number N =2,k =3 of the topographic background images, that is, the color values and the corresponding area ratios of the dominant colors of the background are respectively
Substituting the background dominant color of the background image 2 according to the above optimization process
All permutations of (a), the variable with the least difference in colour: k1' =2, k2' =3, k3' =1, i.e., k1=2, k2=3, k3=1;
the background main color combination with the minimum overall color difference in the embodiment of the invention is as follows:
the multi-terrain camouflage color dominant color is obtained by calculating the mean value of each background dominant color combination s And area ratio p s S ∈ {1, 2., K }, the formula is expressed as follows:
from the above formula, the ratio of the dominant color and the area of the multi-terrain camouflage color in the embodiment of the present invention is shown in the following table 2, for example:
TABLE 2 Multi-terrain camouflage mass tone and area ratio
The multi-terrain camouflage dominant color in the embodiment is shown in fig. 4;
the multi-background camouflage color dominant color and area ratio calculation method is not limited to two terrain backgrounds, and is also suitable for dominant color extraction of more than two terrain backgrounds.
S2, designing a patch library;
based on the bionics principle, a typical biological shape image of a background is collected, the typical biological shape image is preprocessed and trimmed, and then the biological shape image is converted into a biological information plaque by using a matrixing method;
the algorithm is described as follows:
step 1: inputting a biological shape image A m M is equal to {1,2, 3.,. M }, M is the total number of the biological shape images to be processed, the human-computer interactive frame selects the minimum bounding rectangle containing all the biological shapes, and the rectangle is output and named as A _ S m ；
Step 2: for A _ S m Carrying out size normalization processing, setting the size L x L, namely the size of the biological information patch, then carrying out decoloring processing, and converting an RGB color image into a gray image;
and step 3: performing image self-adaptive threshold segmentation algorithm on the gray level image to segment out the biological shape contour, and recording the segmentation result as A _ cut m Convert A _ cut m Is a binary matrix, namely the biological information plaque, and is marked as A _ result m A _ result is added m Sequentially storing the data into a plaque library A _ CLUB;
and 4, step 4: circularly executing the step 1-3 until all the M images containing the biological shape information are processed and stored in the plaque library, and ending the circular process;
the typical biological shape image of the background is an animal and plant shape image collected in a plurality of types of terrain backgrounds where the ground equipment to be camouflaged is located;
in the embodiment of the invention, the biological shape image is a willow leaf, poplar leaf, maple leaf and tiger stripe shape image;
the camouflage patch is manufactured according to the biological shape image, so that the process of generating the patch by processing image patch segmentation, extraction and the like on each terrain background image is avoided, the calculation simplicity is improved, and the bionic source basis of the camouflage is met.
The patch library is only required to be manufactured once and is directly called in the subsequent camouflage color generation process;
a flow chart of a specific plaque library design algorithm is shown in FIG. 5;
in this embodiment, patch library design is performed on images of willow leaves, poplar leaves, maple leaves, and tiger stripes, that is, M =4, the size L =10 × 10 (pixels) of the biological information patch is set, and the segmentation result a _ cut is obtained m And a biological information patch A _ result m As shown in FIGS. 6 (a), (b), (c), (d), respectively;
s3, generating deformation and digital camouflage;
s3.1, generating a dominant color pattern;
firstly, dividing the size of the dominant color pattern and the size of the mosaic unit to obtain the number of the mosaic unit in the dominant color pattern, wherein the formula is as follows:
wherein o x q is the size of the dominant color pattern, and λ x λ is the size of the mosaic unit;
then the number n of mosaic units of each multi-terrain camouflage color dominant color is distributed according to the area proportion calculated in the step S1.3 r ，r∈{1,2,...,K}；
Finally, randomly selecting n in the dominant color pattern r A mosaic unitFilling the corresponding multi-terrain camouflage color dominant color s Until the mosaic units are completely filled, the main color pattern X is generated;
in this embodiment, o × q =300 × λ =5 × 5 is set, the unit is a pixel, and by substituting the above formula, n =3600 is obtained, and the number of mosaic cells for each multi-terrain dominant color can be assigned according to the area ratio of the multi-terrain dominant color: n is a radical of an alkyl radical 1 ＝3600×0.4929＝1774、n 2 ＝3600×0.3147＝1133、n 3 ＝3600×0.1924＝693；
The dominant color pattern generated in the present embodiment is shown in fig. 7;
s3.2, generating a deformed camouflage pattern;
after the dominant color pattern is generated, a biological information plaque A _ result is randomly selected from the plaque library m And fusing the plaque characteristics into the dominant color pattern by a mathematical morphology method, and performing a specific fusion process by adopting an open-close operation formula, wherein the formula is expressed as follows:
I_com＝X·(XοA_result m )
i _ com represents the generated deformed camouflage pattern, an operator represents morphological open operation, and an operator represents morphological close operation;
the plaque size in the changeable camouflage pattern of accessible control biological information plaque size L and mosaic unit size lambda's ratio is nimble to be controlled to the integration in-process, considers changeable camouflage pattern in the size of plaque can not be too big, generally sets up mosaic unit size and is less than biological information plaque size, but the size setting can refer to: 0.5 x L is not more than lambda and is less than L;
if the blended camouflage patterns have small spots and noise points, smoothing can be continuously carried out through morphological closed operation, and the small spots are combined, so that the generated camouflage patterns are more reasonable and natural on the basis of keeping the bionic texture, and the morphological closed operation formula is as follows:
I_com1＝I_com·se
in the formula, se is a structural element,
i _ com1 is a smoothed deformed camouflage pattern;
in the embodiment of the invention, willow leaf patches, tiger stripes and maple leaf patches in a patch library are respectively selected to carry out fusion operation, and the generated deformed camouflage patterns are shown in figures 8 (a), (b) and (c);
s3.3, generating a digital camouflage pattern;
generating a digital camouflage pattern through digital processing after the deformed camouflage pattern is generated;
firstly, determining the size d of a basic unit of the digital camouflage color, then, processing the deformed camouflage color pattern in blocks, wherein the size of each block is the size of the basic unit of the digital camouflage color, and the processing process of the blocks is as follows:
counting the pixels with the most occurrence times in the block and assigning the color values of the pixels to all the pixels in the block;
the operation is circulated until all blocks in the deformed camouflage pattern are completely executed, and the processed image is the digital camouflage pattern;
the statistical blocking processing process adopts sequential operation, namely from left to right and from top to bottom in an image matrix;
the most frequent pixels in the statistical block follow the following rules:
(1) If only one pixel exists in the block with the largest occurrence frequency, replacing all pixels in the block with the pixel with the largest occurrence frequency;
(2) If the pixels with the most occurrence times in the block are G types, and G belongs to {2, 3., K }, one of the G types of pixels is randomly selected to replace all the pixels in the block.
In the embodiment of the present invention, the size d × d =3 × 3 (pixels) of the basic unit of the digital camouflage is set, and the deformed camouflage digitization results of the willow leaf, tiger stripe, and maple leaf patches are shown in fig. 9 (a) and (b).
According to the method, deformation and digital camouflage patterns suitable for a plurality of terrain background environments are generated through the steps of multi-terrain camouflage main color extraction, bionic patch library design, random camouflage pattern generation and the like, the method can provide reference for camouflage pattern design of various ground equipment, and is also suitable for camouflage pattern design of other static military targets under different terrain backgrounds.
The invention discloses a multi-terrain camouflage pattern generating method suitable for ground equipment, which comprises the following steps: s1, dominant color extraction: selecting a typical terrain background image of a region related to the equipment to be camouflaged, and extracting and calculating the multi-terrain camouflage color and the corresponding area proportion thereof; s2, designing a plaque library: based on the bionics principle, acquiring a background typical biological shape image, and constructing a camouflage patch library through matrixing processing; s3, camouflage pattern generation: and (3) randomly filling the multi-terrain camouflage color and the area ratio thereof in the step (S1) to generate a main color pattern, selecting a biological information patch from the camouflage patch library in the step (S2) to blend the main color pattern to generate a deformed camouflage color, and performing digital processing to obtain the digital camouflage color. The method is simple and quick in calculation, the generated camouflage pattern texture is high in randomness, the size of the camouflage spots is controllable, and the method is suitable for multiple terrain backgrounds. The method and the technical thought of the invention promote the development of the camouflage theory.
The invention is not the best known technology.
The above embodiments are merely illustrative of the technical ideas and features of the present invention, and the purpose thereof is to enable those skilled in the art to understand the contents of the present invention and implement the present invention, and not to limit the protection scope of the present invention. All equivalent changes and modifications made according to the spirit of the present invention should be covered within the protection scope of the present invention.
Claims (5)
1. A multi-terrain camouflage pattern generation method suitable for ground equipment is characterized by comprising the following steps:
s1, extracting a dominant color;
s1.1, analyzing the terrain environment of the ground equipment to be disguised, collecting a plurality of terrain background images, selecting a typical image for each terrain background, and storing the typical image into a sequence P, wherein P = { I = i I =1,2,. So, N }, N denoting the choiceThe number of background images; preprocessing the selected background image, including size normalization and image denoising;
s1.2, adopting an image color quantization algorithm to carry out I treatment on the preprocessed background image sequence i And respectively carrying out color quantization processing, quantizing the color data to K background dominant colors, recording the color values and the corresponding area proportions of the background dominant colors, and recording as follows:
k∈(1,2,...,K)，f i k a color value representing the dominant color of the background is represented by f i k [R]、f i k [G]、f i k [B]The three-dimensional vector of the composition,
representing an area ratio of the background dominant color;
s1.3. For different background images I i Background dominant color f i k Performing color difference analysis, and calculating the main color of the multi-terrain camouflage color and the corresponding area proportion thereof according to the analysis result;
the average color difference formula is used as a measurement function, the color difference between the background dominant colors is analyzed, and the smaller the measurement function value is, the smaller the difference between the two background dominant colors is;
setting and selecting two different topographic background images I i 、I j After the color quantization processing of step S1.2, the color value and the corresponding area ratio of the background dominant color are respectively
The metric function formula can be derived as follows:
finding the background dominant color combination which minimizes the overall color difference according to the definition of the color similarity metric function,
the optimization process formula is described as follows:
the above formula represents substituting into the background image I j All permutation combinations of the background dominant colors are searched for variables k1', k2', k3',. K, k' which minimize the overall calculation result, namely I is carried out i 1,2,3, K background dominant colors and I j K1, k2, k3,. K background dominant color combinations;
wherein K1', K2', K3', K' is in the scope of {1,2, 3.., K }, and K represents the number of multi-background camouflage dominant colors;
according to the above optimization process, the background dominant color combination with the minimum overall color difference is:
the multi-terrain camouflage color dominant color is obtained by calculating the mean value of each background dominant color combination s And area ratio p s S ∈ {1, 2., K }, the formula is expressed as follows:
s2, designing a patch library;
based on the principle of bionics, firstly, a typical biological shape image of a background is collected, the typical biological shape image is preprocessed and trimmed, and then the biological shape image is converted into a biological information patch by using a matrixing method;
the specific process of designing the plaque library is described as follows:
step 1: input a sheetBiological shape image A m M is equal to {1,2, 3.,. M }, M is the total number of the biological shape images to be processed, the human-computer interactive frame selects the minimum bounding rectangle containing all the biological shapes, and the rectangle is output and named as A _ S m ；
And 2, step: for A _ S m Performing size normalization processing, setting the size to be L x L, wherein the size is the size of the biological information patch, then performing color removal processing, and converting an RGB color image into a gray image;
and step 3: performing image self-adaptive threshold segmentation algorithm on the gray level image to segment out the biological shape contour, and recording the segmentation result as A _ cut m Convert A _ cut m Is a binary matrix, namely the biological information plaque, and is marked as A _ result m A _ result is added m Sequentially storing the data into a plaque library A _ CLUB;
and 4, step 4: circularly executing the step 1-3 until all the M images containing the biological shape information are processed and stored in the plaque library, and ending the circular process;
s3, generating camouflage patterns;
s3.1, generating a dominant color pattern;
firstly, dividing the size of the dominant color pattern and the size of the mosaic unit to obtain the number of the mosaic unit in the dominant color pattern, wherein the formula is as follows:
in the formula, o x q is the size of the primary color pattern, and lambda x lambda is the size of the mosaic unit;
then distributing the number n of mosaic units of each multi-terrain camouflage color dominant color according to the area proportion calculated in the step S1.3 r ，r∈{1,2,...,K}；
Finally, randomly selecting n in the dominant color pattern r A mosaic unit filled to the corresponding multi-terrain camouflage main color s Until all the mosaic units are filled, the main color pattern X is generated;
s3.2, generating a deformed camouflage pattern;
after the dominant color pattern is generated, a biological information plaque A _ result is randomly selected from the plaque library m And fusing the plaque characteristics into the dominant color pattern by a mathematical morphology method, and performing a specific fusion process by adopting an open-close operation formula, wherein the formula is expressed as follows:
i _ com represents the generated deformed camouflage pattern, operator. Representing a morphological open operation, an operator representing a morphological closed operation;
s3.3, generating a digital camouflage pattern;
generating a digital camouflage pattern through digital processing after the deformed camouflage pattern is generated;
firstly, determining the size d of a basic unit of the digital camouflage, then processing the deformed camouflage pattern in blocks, wherein the size of each block is the size of the basic unit of the digital camouflage, and the processing process of the blocks is as follows:
counting the pixel with the most occurrence times in the block and assigning the color value of the pixel to all the pixels in the block;
and circulating the operations until all the blocks in the deformed camouflage pattern are completely executed, wherein the processed image is the digital camouflage pattern.
2. The method of claim 1, wherein the method comprises: the method for calculating the multi-background camouflage painting dominant color and the area ratio in the step S1.3 is not limited to two different terrain backgrounds, and is also suitable for extracting the dominant color of more than two terrain backgrounds.
3. The method of claim 1, wherein the method comprises: step S3.2 can flexibly control the size of the patch in the deformed camouflage pattern by controlling the ratio of the size L of the biological information patch to the size lambda of the mosaic unit in the fusion process, and the size of the mosaic unit is generally set to be smaller than that of the biological information patch in consideration of the fact that the size of the patch in the deformed camouflage pattern cannot be too large, and the size setting can refer to: 0.5 x L is not more than lambda and is less than L.
4. The method of claim 1, wherein the method comprises: step S3.3 in the process of blocking, the pixels with the largest number of occurrences in the statistical block follow the following rules:
(1) If only one pixel exists in the block with the largest occurrence frequency, replacing all pixels in the block with the pixel with the largest occurrence frequency;
(2) If the pixels with the most occurrence times in the block have G types, and G belongs to {2, 3., K }, one of the G types of pixels is randomly selected to replace all the pixels in the block.
5. The method of claim 4, wherein the method comprises: the blocking process uses sequential operations, i.e., from left to right, top to bottom, in the image matrix.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202110225866.2A CN112950733B (en) | 2021-03-01 | 2021-03-01 | Multi-terrain camouflage pattern generation method suitable for ground equipment |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202110225866.2A CN112950733B (en) | 2021-03-01 | 2021-03-01 | Multi-terrain camouflage pattern generation method suitable for ground equipment |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN112950733A CN112950733A (en) | 2021-06-11 |
| CN112950733B true CN112950733B (en) | 2022-10-28 |
Family
ID=76246884
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202110225866.2A Active CN112950733B (en) | 2021-03-01 | 2021-03-01 | Multi-terrain camouflage pattern generation method suitable for ground equipment |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN112950733B (en) |
Citations (8)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN102778176A (en) * | 2011-05-11 | 2012-11-14 | 贵州航天凯山特种车改装有限公司 | Mosaic variegation digital camouflage pattern setting on surface of shelter |
| CN107392880A (en) * | 2017-07-25 | 2017-11-24 | 北京华新创科信息技术有限公司 | A kind of imitative pattern painting automatic generation method |
| CN108510562A (en) * | 2018-02-11 | 2018-09-07 | 青岛九维华盾科技研究院有限公司 | Digital camouflage method for generating pattern based on image fractal texture |
| CN110021054A (en) * | 2019-04-12 | 2019-07-16 | 青岛九维华盾科技研究院有限公司 | A kind of patch colouration method for the design of spot camouflage pattern |
| CN110120080A (en) * | 2019-04-12 | 2019-08-13 | 青岛九维华盾科技研究院有限公司 | A method of quickly generating standard pattern-painting mass-tone |
| CN110660116A (en) * | 2019-09-19 | 2020-01-07 | 北京工业大学 | Method for generating digital camouflage from deformed camouflage |
| CN111798539A (en) * | 2020-01-10 | 2020-10-20 | 中国人民解放军国防科技大学 | Adaptive camouflage online design method and system |
| CN112184838A (en) * | 2020-10-09 | 2021-01-05 | 哈尔滨工程大学 | Multi-background camouflage pattern dominant color extraction method based on color correlation |
Family Cites Families (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US20150290967A1 (en) * | 2014-04-10 | 2015-10-15 | Phantom Design, Llc | Camouflage Design and Method |
| US10430979B2 (en) * | 2018-02-28 | 2019-10-01 | Nate Turner | Method for creating camouflage patterns |
-
2021
- 2021-03-01 CN CN202110225866.2A patent/CN112950733B/en active Active
Patent Citations (8)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN102778176A (en) * | 2011-05-11 | 2012-11-14 | 贵州航天凯山特种车改装有限公司 | Mosaic variegation digital camouflage pattern setting on surface of shelter |
| CN107392880A (en) * | 2017-07-25 | 2017-11-24 | 北京华新创科信息技术有限公司 | A kind of imitative pattern painting automatic generation method |
| CN108510562A (en) * | 2018-02-11 | 2018-09-07 | 青岛九维华盾科技研究院有限公司 | Digital camouflage method for generating pattern based on image fractal texture |
| CN110021054A (en) * | 2019-04-12 | 2019-07-16 | 青岛九维华盾科技研究院有限公司 | A kind of patch colouration method for the design of spot camouflage pattern |
| CN110120080A (en) * | 2019-04-12 | 2019-08-13 | 青岛九维华盾科技研究院有限公司 | A method of quickly generating standard pattern-painting mass-tone |
| CN110660116A (en) * | 2019-09-19 | 2020-01-07 | 北京工业大学 | Method for generating digital camouflage from deformed camouflage |
| CN111798539A (en) * | 2020-01-10 | 2020-10-20 | 中国人民解放军国防科技大学 | Adaptive camouflage online design method and system |
| CN112184838A (en) * | 2020-10-09 | 2021-01-05 | 哈尔滨工程大学 | Multi-background camouflage pattern dominant color extraction method based on color correlation |
Non-Patent Citations (6)
| Title |
|---|
| Research on Digital Camouflage Pattern Generation Algorithm Based on Adversarial Autoencoder Network;Xin Yang 等;《International Journal of Pattern Recognition and Artificial Intelligence》;20201231;第34卷(第6期);第1页 * |
| 伪装迷彩的斑点设计方法研究;杨武侠;《中国优秀硕士学位论文全文数据库 工程科技Ⅱ辑》;20110815(第8期);C032-6 * |
| 变色伪装材料研究关键技术;杜新瑜 等;《第六届中国功能材料及其应用学术会议论文集（8）》;20071101;第168-171页 * |
| 基于分形布朗模型的数码迷彩图案生成方法研究;蔡云骧 等;《兵工学报》;20160115;第37卷(第1期);第186-192页 * |
| 基于分水岭方法的数码迷彩设计;白万民等;《计算机与数字工程》;20120820;第40卷(第08期);第116-119页 * |
| 迷彩图案自动化生成系统方案的研究与实现;辛淙;《中国优秀硕士学位论文全文数据库 信息科技辑》;20170315(第3期);I138-5362 * |
Also Published As
| Publication number | Publication date |
|---|---|
| CN112950733A (en) | 2021-06-11 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| Enomoto et al. | Filmy cloud removal on satellite imagery with multispectral conditional generative adversarial nets | |
| Suárez et al. | Infrared image colorization based on a triplet dcgan architecture | |
| CN104318570B (en) | A kind of self adaptation camouflage painting method for designing based on background | |
| US8194975B2 (en) | Use of an intrinsic image in face recognition | |
| Li et al. | Spectrum characteristics preserved visible and near-infrared image fusion algorithm | |
| CN108319973A (en) | Detection method for citrus fruits on tree | |
| Suárez et al. | Deep learning based single image dehazing | |
| CN111798539A (en) | Adaptive camouflage online design method and system | |
| CN110111346B (en) | Remote sensing image semantic segmentation method based on parallax information | |
| Hendrawan et al. | Development of colour co-occurrence matrix (CCM) texture analysis for biosensing | |
| Jin et al. | Semisupervised remote sensing image fusion using multiscale conditional generative adversarial network with siamese structure | |
| Bian et al. | Fuzzy c-means clustering based digital camouflage pattern design and its evaluation | |
| Yang et al. | Research on digital camouflage pattern generation algorithm based on adversarial autoencoder network | |
| Liu et al. | Color image edge detection method based on the improved whale optimization algorithm | |
| CN112950733B (en) | Multi-terrain camouflage pattern generation method suitable for ground equipment | |
| Liu et al. | TSCMDL: Multimodal deep learning framework for classifying tree species using fusion of 2-D and 3-D features | |
| Zheng et al. | Coloring night-vision imagery with statistical properties of natural colors by using image segmentation and histogram matching | |
| Katırcıoğlu et al. | A novel gray image enhancement using the regional similarity transformation function and dragonfly algorithm | |
| Peli et al. | Multispectral image fusion for visual display | |
| Rasras et al. | Developing a new color model for image analysis and processing | |
| CN112184838B (en) | Multi-background camouflage pattern dominant color extraction method based on color correlation | |
| Gurung et al. | Image segmentation using multi-threshold technique by histogram sampling | |
| Shen et al. | A holistic image segmentation framework for cloud detection and extraction | |
| Fang et al. | Research on Multi-Background Camouflage Pattern Design Method Based on Feature Fusion | |
| Wu et al. | A Review of Camouflaged Target Detection Research |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| PB01 | Publication | ||
| PB01 | Publication | ||
| SE01 | Entry into force of request for substantive examination | ||
| SE01 | Entry into force of request for substantive examination | ||
| GR01 | Patent grant | ||
| GR01 | Patent grant |