基于超分辨率纹理重构的数码迷彩设计研究
Research on Digital Camouflage Design Method Based on Super-resolution Texture Reconstruction
吕琪 1李凌 2杨鑫 2张飞 1李浩 2朱锐3
作者信息
- 1. 陆军工程大学野战工程学院,江苏南京 210000;61606部队,北京 100080
- 2. 陆军工程大学野战工程学院,江苏南京 210000
- 3. 61606部队,北京 100080
- 折叠
摘要
为实现数码迷彩设计与背景纹理特性精准化结合目的,提出了一种基于超分辨率纹理重构算法的数码迷彩设计方法.首先利用模糊聚类法提取了背景主色,再使用Criminis算法生成数码迷彩使边缘分割更细碎;最后改进的超分辨率纹理重构算法,将背景图像上的纹理细节放大至迷彩图案中,使数码迷彩图像的纹理适应性更强.针对戈壁、丛林两种典型背景图像进行数码迷彩设计,结果表明使用该模型生成的数码迷彩纹理适应性较强.
Abstract
In order to realize the precise and automatic combination of digital camouflage design and background texture characteristics,a digital camouflage design method based on super-resolution texture reconstruction algorithm is proposed.Firstly,the fuzzy clustering method was used to extract the main color of the background,and then the Criminis algorithm was used to generate digital camouflage to make the edge segmentation finer;The final improved super-resolution texture reconstruction algorithm amplifies the texture details on the background image into the camouflage pattern,making the texture adaptability of digital camouflage images stronger.Digital camouflage de-sign was carried out for two typical background images,Gobi and Jungle,and the results showed that the digital camouflage texture gener-ated by this model has strong adaptability.
关键词
数码迷彩设计/模糊聚类/Criminis算法/超分辨率纹理重构/纹理适应性Key words
Digital camouflage design/Fuzzy clustering/Criminis algorithm/Super-resolution reconstruction/Texture adaptability引用本文复制引用
出版年
2024
NETL
NSTL
万方数据