融合Vibe前景检测的Camshift算法  

Improved Camshift Algorithm Fused with Vibe Foreground Detection

作　　者：曾志宏[1] 陈振武[1] 肖庆云[2] 

机构地区：[1]龙岩学院,福建龙岩364000 [2]厦门大学,福建厦门361008

出　　处：《龙岩学院学报》2017年第2期89-94,共6页Journal of Longyan University

基　　金：福建省自然科学基金资助项目(2015J01587);福建省教育厅中青年项目(JAT160487);龙岩学院服务海西基金资助项目(JB10160;LYXY2011067)

摘　　要：Camshift算法是一种以跟踪颜色信息为目标的算法,该算法由于在光照变化、相似背景颜色干扰及目标遮挡的环境下鲁棒性不高,从而易造成视频跟踪的错误。为解决这些问题,将基于Vibe前景检测方法获取的前景信息融入到Camshift算法中,通过这种改进后的Camshift算法可以增强前景和背景的区分度。通过不同场景的视频跟踪结果表明,改进后的Camshift算法能更有效克服原算法的不足,具有较强的鲁棒性。Camshift is a color-based tracking algorithm. Under illumination variation, similar background interference and target occlusion, the Camshift algorithm has low robustness and is easy to track astray. For solving this problem, an improved Camshifl algorithm fused with the Vibe algorithm can enhance distinction degree between target and background. Video target tracking results of different scenes show that the improved algorithm can effectively overcome the disadvantages of the traditional Camshift algorithm, such as illumination variation, similar background interference and target occlusion while the improved algorithm has higher robustness.

关 键 词：CAMSHIFT算法 目标跟踪 光照变化 颜色干扰 目标遮挡 

分 类 号：TP18[自动化与计算机技术—控制理论与控制工程]