该技术用修正的最大类间方差法分割亚图像,聚类并去除强背景干扰区域,对可能目标区域能量进行统计,选

择区域能量最大者进行逐帧实时锁定跟踪,并对该技术的实时性进行了分析。

- The image is segmented and fallen into

two classes by the modifiedmethod of maximal variance of classes, then regions with intense background interference are discarded

and the rest regions’energy is added up respectively. The regionwiththemaximal energy is locked asthe target and tracked in real

time. The author also analyzes the real-time performance of this algorithm.