摘要
针对传统运动目标检测方法帧差法、光流法等提取运动目标的缺陷与不足,提出了一种新的基于冗余小波变换和DT网格的运动目标检测算法,即在冗余小波域提取特征点生成Delaunay三角形网格,并自适应获得潜在运动区域,从而获得运动目标区域实现运动目标检测.实验结果表明,本文提出的算法可以有效地提取运动目标,效果要好于传统的运动目标检测方法,对于室外交通监控和室内监控的运动目标检测同样适用,取得不错的效果.
An object detection method which can be used in indoor and outdoor monitoring is proposed in this paper.Improving shortcomings and disadvantages of moving object extraction in traditional inter-frame difference method,it puts forward a moving object recognition algorithm based on redundant wavelet transform,namely,extraction of feature points in redundant wavelet domain to generate a delaunay triangulation grid and adaptive potential movement area,then moving object area.The experimental results show that the proposed algorithm can effectively extract moving object,which is better than traditional frame difference method.
出处
《河北工业大学学报》
CAS
北大核心
2013年第2期20-23,共4页
Journal of Hebei University of Technology
基金
河北省自然科学基金重点项目(ZD200911)
关键词
冗余小波变换
目标检测
视频挖掘
DT网格
特征点
redundant wavelet transform
object detection
video mining
Delaunay Triangle grid
feature points
