基于分形布朗模型的数码迷彩图案生成方法研究

Study of Digital Pattern Generation Method Based on Fractal Brown Model

迷彩伪装是一项重要的伪装技术。数码迷彩是对抗高分辨成像侦察的有效技术手段。如何实现迷彩设计技术的实效化、精确化和智能化始终是研究热点。基于分形布朗运动的分形维估计法提取背景图像的纹理特征、基于金字塔结构逐层模糊C均值聚类算法提取背景图像的主色特征。融合纹理和主色特征,运用改进的钻石-菱形算法生成符合背景特征的数码迷彩图案。仿真实验结果表明:该算法在荒漠型、丛林型、林地型背景的迷彩生成中均能产生较好效果。

Painting camouflage is an important camouflage technique. Digital camouflage is an effective means against high-resolution imaging reconnaissance. How to achieve the effectiveness, precision and intelligence of camouflage design technology is always a reseach focus. The texture features of background image are extracted based on FBM fractal dimension estimation method, and the dominant color features of background image are extracted based on pyramid structure fuzzy C-means cluster （FCM） clustering algorithm. The digital camouflage patterns conforming to the characteristics of background are generated based on the improved diamond-square algorithm. The simulation results show that the proposed algorithm is used to achieve the preferable results of pattern generation in desert, jungle, forest land backgroiknds.