本科毕业论文—运动目标检测与跟踪技术研究（设计）.docVIP

摘 要 运动目标检测与跟踪是计算机视觉领域的热点问题，广泛应用于视频监控、人机交互、虚拟现实和图像压缩等。而要在各种复杂的环境中和不同的条件下如遮挡、光照变化等都对目标进行准确的跟踪是目前广大研究工作者共同关注的焦点，也是目前实际应用中一个亟待解决的难题。 本文主要研究静态背景下运动目标的检测、运动目标跟踪以及相关结果仿真分析三方面的内容。运动目标检测方面，分析了目前比较常用的三种目标检测方法，即光流法、帧间差分法背景相减法，着重研究了基于帧间差分法运动目标检测的算法原理及流程讨论了三种检测算法的优缺点。运动目标跟踪方面，目前比较常用的三种目标跟踪算法，即均值漂移算法、卡尔曼滤波算法基于特征的目标跟踪算法，重点研究了基于特征最小外接矩形框运动目标跟踪。分析了其算法原理以及跟踪步骤。最后用matlab对运动目标的视频 关键词：运动目标检测；运动目标跟踪；帧间差分法；最小外接矩形框Abstract Moving target detection and tracking is a hot issue in the field of computer vision, is widely used in video surveillance, human-computer interaction, virtual reality and image compression etc.. But in various complex environments and different conditions (such as occlusion, illumination changes) of target accurate tracking is currently the focus of researchers of common concern in the current application, is an urgent problem to be solved. This paper mainly involves moving target detection under a static background, moving target tracking and simulation results analysis of three aspects. In the moving object detection, in the analysis of the current commonly used three kinds of target detection method based on optical flow method, namely, the inter-frame difference method, background subtraction method, focusing on the frame difference algorithm principle and process method based on moving object detection. And discusses the advantages and disadvantages of three kinds of detection algorithm. In terms of moving object tracking, in the analysis of the current three kinds of target more commonly used tracking algorithm, the mean shift algorithm, Calman algorithm, based on the characteristics of the target tracking algorithm based on the characteristics, focuses on the minimum exterior rectangle based on moving target tracking algorithm. Finally using MATLAB software using frame difference detection method for moving targets as well as based on the minimum bounding box tracking method to simulate the video with moving objects. Based on the frame difference detection met