基于卡尔曼滤波法视频目标跟踪的改进  

Improvement of Video Target Tracking Based on Kalman Filtering Method

作　　者：梁成[1] 

机构地区：[1]西安医学院公共课部,西安710021

出　　处：《计算机与数字工程》2014年第5期809-811,918,共4页Computer & Digital Engineering

摘　　要：基于视频的目标跟踪是交通事件检测的一种重要方法,对视频中的目标物提取之后,再对目标物跟踪是现代基于视频交通事件检测的重要手段。目标区域的跟踪主要工作是对目标物特征点的提取,并对目标物特征点进行跟踪。基于卡尔曼滤波法目标特征点的跟踪方法简单。但传统的卡尔曼滤波法计算量大,论文是在基于卡尔曼滤波的基础上进行改进和方法优化。目标物特征点跟踪的方法主要采用基于卡尔曼滤波法的模版匹配使目标跟踪准确、计算量小。Target tracking in video is an important method of traffic incident detection through the extraction of the tar- get object. The tracking of target area can be simply realized by extraction of target feature points and their tracking. The traditional kalman filtering is a method of complex and huge calculation, yet the method is improved a lot in this paper com- pared with the traditional kalman filter. The target feature points tracking based on the template matching in kalman filtering method, thus making the higher accuracy of target tracking and small amount of calculation.

关 键 词：卡尔曼滤波法 跟踪 特征点 

分 类 号：TP391[自动化与计算机技术—计算机应用技术]