摘要
研究利用背景移动补偿方法消除运动图像相邻帧相减造成的背景杂波 ,从而把复杂背景下运动目标的识别问题转化为简单背景下的目标识别问题 .利用背景中特征较为明显的区域进行模板匹配 ,得到背景的运动数据 ,依据此数据将相邻帧图像错位相减 .将差分图像中的背景杂波去掉 ,只剩下运动物体的边缘和高梯度区域 ,有利于目标的识别 .
A new method of solving the problem of camera moving, based on difference image sequence processing, was proposed, with which the complex backgrounds were converted into relatively simple ones. The data of the background moving can be obtained by template matching using obvious features of the background and transmitted to difference image sequence processing part. By this way, the edge of background is removed and only the edge of the moving target is left in difference image sequence. As proved by experiments, this method is effective.
出处
《北京理工大学学报》
EI
CAS
CSCD
2000年第3期333-337,共5页
Transactions of Beijing Institute of Technology
基金
国家"八五"攻关项目
关键词
背景移动补偿
运动目标检测
目标识别
帧图像
background moving compensation
moving target detection
target recognition
