背景典型斑点特征的聚类分析

迷彩是一项重要的目标防护手段,斑点形状设计是进行迷彩设计的基础。首先自动提取背景图像中的明显斑点,然后分析背景斑点周长、面积、伸长度、矩形度、细长度、分散度等特征的统计特性,最后将长度、矩形度、细长度、分散度作为斑点特征向量对提取出的背景斑点进行HCM聚类分析,进而得到背景典型斑点的特征参数。该方法能得到大样本背景图像斑点特征的统计特征,对自动、定量化的迷彩设计具有一定意义。

Character of Representative Spots from Background by Using Cluster Analysis

The pattern painting is an important method to protect the target, the pattern painting design is based on the spot shape design. First, the visible spots are extracted automatically from the background image. Then, the statistic characteristics of the girth, area, elongation ratio, rectangular ratio, slightness ratio and dispersing ratio for the background spot are analyzed. Finally, taking the length, rectangular ratio, slightness ratio and dispersing ratio as the spot characteristic vector, the extracted background spot is analyzed by using HCM cluster analysis, and the characteristic parameter is gained. This method can gain the statistic characteristics of the large sample background images, which is useful to for the design of the automatic and quantitative pattern painting.