基于OpenCV的运动目标检测与跟踪  

Detection and tracking of moving objects with OpenCV

作　　者：张学贺[1] 张学东[1] 丁宁[1] 

机构地区：[1]辽宁科技大学电子与信息工程学院,辽宁鞍山114051

出　　处：《辽宁科技大学学报》2010年第5期490-494,共5页Journal of University of Science and Technology Liaoning

摘　　要：提出了一种基于OpenCV的视频应用程序的开发方法。将Intel公司开放的OpenCV源代码作为开发的视频应用程序的基础函数库,改写或调用其中的函数,可根据需要用C++语言开发视频应用程序,从而克服了视频应用程序开发周期长、效率低的缺点。给出了部分OpenCV具体的设置方法和过程,并通过利用OpenCV编写的运动目标检测与跟踪的应用程序验证了该方法的有效性和可行性。对图像进行形态学去噪、平滑滤波处理和二值化阈值分割等图像预处理,获得二值化黑白图像。通过轮廓提取和跟踪检测得到车辆和人的外轮廓,进而实现运动目标的跟踪。在VC++6.0环境下,利用OpenCV编程实现了此方法。实验结果表明,该方法可行。A method for developing video application program is proposed based on Intel′s open source computer vision library（OpenCV） in this paper.The OpenCV can be used as a basic function library of developing video application program.By rewriting or using functions of OpenCV,video application program can be developed by writing C programming language.This method can solve the disadvantage of long period and low efficiency for developing video application program.The approach and process of setting is provided in this paper.An application program on camera calibration using OpenCV shows that this method is effective and feasible.The image is preprocessed by denoising with mathematical morphology,smoothing and binary threshold segmentation to obtain the binary image on the basis of black and white.The tracking of moving objects is realized by the contour extraction and tracking detection.The method is realized by using OpenCV in VC＋＋ 6.0 programming environment.Experimental results show that this method is feasible.

关 键 词：计算机视觉 图像处理 运动检测 目标跟踪 

分 类 号：TP391[自动化与计算机技术—计算机应用技术]