摘要
提出了一种新型图像多目标识别跟踪算法,其突出特点为速度快、目标探测准确性高,并且可以同时对多个目标进行识别跟踪.整个系统包括图像预处理即目标增强,边界的目标识别即纵向、横向统计叠加和目标检测.传统的图像跟踪方式如模板相关匹配方法、梯度形心方法以及极值区域跟踪方法不能同时区分多个目标,遇到多目标和干扰等均会失效。文中提出的方法能够有效地解决这一问题,通过区分可能目标区域,能够在一定程度上抑制干扰并对多个目标进行实时跟踪。
In modern intelligent weapon systems,the target homing,especially the ability of multi-target tracking is one of the most important aspects.In this paper,an improved method of multi-target homing method is discussed.The new way of dealing with multi-target has the advantage of high detecting,quality of multi-target and speed.Compared with the new method,the traditional ways don't work well with dividing multi-target areas at the same time and still would fail when facing multi-target and background noise.The new way of finding multi-target in the image will solve the problem in effectively;through the possibly area estimation,disturbance can be restrained to some extent,and the multi-target identification is performed better.
出处
《红外与激光工程》
EI
CSCD
北大核心
2006年第z4期248-250,共3页
Infrared and Laser Engineering
关键词
多目标
识别
跟踪
简单背景
Multi-target
Identification
Homing
Simple background
