计算机应用
計算機應用
계산궤응용
COMPUTER APPLICATION
2014年
z2期
286-290
,共5页
董慧芬%董保磊%丁小芳%张振
董慧芬%董保磊%丁小芳%張振
동혜분%동보뢰%정소방%장진
相关区域分层%Meanshift%自适应Kalman滤波算法%目标跟踪%匹配算法%相似函数
相關區域分層%Meanshift%自適應Kalman濾波算法%目標跟蹤%匹配算法%相似函數
상관구역분층%Meanshift%자괄응Kalman려파산법%목표근종%필배산법%상사함수
relevant region stratificaiton%Meanshift%adaptive Kalman filtering algorithm%target tracking%matching algorithm%similar function
针对背景图像阴影、物体遮挡等对目标跟踪干扰的问题，提出一种基于相关区域分层的改进Meanshift目标跟踪算法。Meanshift跟踪算法搜索匹配问题，采用一种相关区域分层匹配搜索方法，提高在跟踪过程中重新搜索匹配的快速性，并结合自适应Kalman滤波算法提高目标跟踪的准确性。实验结果表明了该算法的可行性及有效性。
針對揹景圖像陰影、物體遮擋等對目標跟蹤榦擾的問題，提齣一種基于相關區域分層的改進Meanshift目標跟蹤算法。Meanshift跟蹤算法搜索匹配問題，採用一種相關區域分層匹配搜索方法，提高在跟蹤過程中重新搜索匹配的快速性，併結閤自適應Kalman濾波算法提高目標跟蹤的準確性。實驗結果錶明瞭該算法的可行性及有效性。
침대배경도상음영、물체차당등대목표근종간우적문제，제출일충기우상관구역분층적개진Meanshift목표근종산법。Meanshift근종산법수색필배문제，채용일충상관구역분층필배수색방법，제고재근종과정중중신수색필배적쾌속성，병결합자괄응Kalman려파산법제고목표근종적준학성。실험결과표명료해산법적가행성급유효성。
Concerning the interference questions of the shadows in the background images and the object occlusion aiming at video surveillance, this paper presented an improved Meanshift target tracking algorithm based on stratification of relevant regions. The matching search of Meanshift tracking algorithm was analyzed: matching search method based on stratification of relevant regions was used to improve operational speed; meanwhile, adaptive Kalman filtering algorithm was adopted to improve the accuracy of following tracking. Experimental results of pedestrian tracking show the feasibility and effectiveness of this algorithm.
