Research On Design Method Of Multi-background Camouflage Pattern

Camouflage is the most commonly used camouflage technology,which is widely used in battlefield camouflage protection.It is an important means to improve the survival and protection of military targets,to counter modern precision-guided weapon attacks,and to reconnaissance and surveillance.Camouflage pattern design is an important part of the implementation of camouflage,which directly determines the effect of camouflage,therefore,the method of camouflage pattern design has always been a hot and difficult problem in the research of camouflage technology at home and abroad.The currently widely used singlebackground camouflage patterns cannot meet the camouflage needs of large areas and multiterrain backgrounds,while the existing multi-background camouflage pattern design methods have less public information,the relevant theoretical system and design rules are not perfect,and the scope is limited to similar multi-background images.In response to the above problems,this thesis has carried out research on the design method of multi-background camouflage patterns,including data analysis and processing of multi-background images,determination of color of multi-background camouflage patterns,generation of multibackground camouflage patterns and evaluation of camouflage effects,etc.,forming a complete set of design method for multi-background camouflage pattern.Firstly,a color determination algorithm for multi-background camouflage patterns based on color similarity is proposed.Based on the regional division,similarity analysis and preprocessing of the input multi-background image data,the color similarity threshold is calculated,and the background dominant color extraction algorithm is optimized and corrected,the color selection of the color candidate set,and the priority ranking of the dominant color,realized the accurate and standardized output of the color of the multibackground camouflage pattern.Secondly,according to the different technical routes,two multi-background camouflage pattern generation methods based on feature fusion and texture synthesis are proposed.The first method constructs a typical multi-background patch library through patch extraction and screening,and then combines the camouflage color and patch shape features through mathematical psychology to achieve the generation of camouflage patterns;the second method uses panoramic image stitching,the texture synthesis based on VGG19 network model and color replacement complete the camouflage pattern generation.Finally,in order to objectively evaluate the camouflage effect of the two types of multibackground camouflage generated in this paper,a method for evaluating the camouflage effect of multi-background camouflage patterns is studied.The texture mapping technology based on Open GL is used to simulate the real camouflage painting to verify the camouflage fusion degree,and the subjective and objective camouflage effect evaluation is carried out based on the model and index method.Experiments show that the two methods proposed in this thesis can effectively design multi-background camouflage patterns and integrate well with each background in terms of brightness,texture and color characteristics.Compared with the traditional single background camouflage pattern,for feature fusion multi background camouflage pattern,the CSI under low similarity and high similarity background is reduced by about 8.1% and 10.8% respectively,and the CSI of texture synthesis multi-background camouflage pattern is reduced by about 7%.The multi-background camouflage pattern design method proposed in this thesis is suitable for the camouflage pattern design of different terrain backgrounds with different degrees of similarity.It provides a set of more reasonable and feasible solutions for the implementation of camouflage design for military targets under multiple backgrounds.The work done in the thesis enriches the theory of camouflage pattern design,which can provide reference for the research of multi-background camouflage technology and related engineering practice.