野外复杂背景下轻武器射击目标的图像识别算法

针对野外复杂背景下轻武器射击目标图像的分析判别,首先采用一种彩色空间聚类分析和阈值分割相结合的混合算法,即基于S、I通道的目标图像两步分割算法,将目标区域从背景图像中分离出来.然后基于机器学习理论中的Boosting算法,利用区域的形状特征和纹理特征等组成一个强分类器,将目标区域从其他区域分离出来,同时判断出目标的类型.

An Image Recognition Algorithm on Small Arms Shooting Target in Complex Field Background

On image recognition of small arms shooting target in complex field background,this paper adopts a hybrid algorithm to separate target area form background image firstly.The algorithm combines the color space clustering analysis and threshold segmentation and is a two-step segmentation algorithm of target image based on S and I channel.Then,based on the boosting algorithm of machine learning theory,a strong classifier is built by using the shape characteristic and texture characteristic of target area.It ca...