基于TMS320DM643的运动目标检测算法  

Moving Target Detection Algorithm Based on TMS320DM643

作　　者：胡小霞[1] 杨余旺[1] 莫然[1] 

机构地区：[1]南京理工大学计算机科学与技术学院,南京210094

出　　处：《计算机工程》2012年第15期179-182,186,共5页Computer Engineering

摘　　要：基于数字视频应用开发平台SEED-DEC643和摄像机FCB-CX1010P,提出一种改进的运动目标检测算法。设计静态背景下单运动目标物体的检测,包括图像预处理、背景差分、阈值求取、图像二值化、图像形态学处理、运动目标物体坐标计算,利用灰度投影算法计算背景运动向量,以实现动态背景下单运动目标物体的检测。实验结果表明,该算法能较好地进行背景运动补偿,并凸显运动物体。Based on the digital video developing platform SEED-DEC643 and the camera FCB-CX1010P,this paper proposes an algorithm of moving target detection.It realizes the single moving target detection under the static background,including image preprocessing,background subtraction,the fetching of threshold,image binarization,morphological image processing and the calculation of the moving target coordinates.This paper utilizes the gray projection algorithm to calculate the motion vectors,then realizes motion compensation and the single moving target detection under the dynamic background.Experimental results show that this algorithm can get well background motion compensation,and highlights the moving targets.

关 键 词：运动目标检测 静态背景 动态背景 背景差分 灰度投影算法 背景运动补偿 

分 类 号：TP391[自动化与计算机技术—计算机应用技术]