基于OpenCV的运动目标跟踪系统研究

本文分析比较了传统运动目标检测的3种主要方法：背景图像差分法、时态差分法和光流法,在此基础上给出了一种背景图像预测算法,大大减少了因为背景变化而产生的目标检测误差。本文基于OpenCV设计出改进的运动目标检测与跟踪算法,实现了运动目标的跟踪,并在VC＋＋编译环境下,利用USB摄像头作为视频采集器,通过观察实验结果可以看出,本文的运动目标检测算法能够正确地检测出视频图像中的运动目标,而且在检测性能上优于普通的自适应背景差分法。

Research and design of motion target tracking system based on the OpenCV

This paper compares and analyzes three traditional method of Moving target detection and tracking：background image difference method,the temporal difference method and the optical flow method,On this basis,we adopt one kind of background images forecast algorithm,greatly reduce the object detection error which produced by the change of background.This article is in the Window platform and the VC ＋ ＋ compiler environment,using USB camera as a video capture device,design the system about moving target detection and tracking based on OpenCV,achieved a moving target tracking by improving the algorithm of Moving target detection and tracking,By observing the experience result we can see,the moving target detection algorithm adoped in this paper can examine the moving target in the video image correctly,moreover this way surpasses the ordinary auto-adaptive background subtraction.in the examination performance.