Information Technology Journal1812-56381812-5646Asian Network for Scientific Information10.3923/itj.2013.614.622FengXin WangXiaoming DangJianwu ShenYu 42013124In order to solve the problem of video target tracking problem, this study proposed a new kind of particle filter for the state estimation of nonlinear system when the statistical characteristics of the system are unknown. The proposed algorithm estimated and corrected the statistic characteristics of the system unknown noise in real-time by improved Sage-Husa estimator and produced optimal distribution function with unscented Kalman Filter. This novel algorithm reduced the estimation error effectively and improved the anti-noise ability of the system. Under the improving particle filter framework, this study used color and motion edge character as observation model and fused multi-features weights through the D-S evidence theory. The experiment results shows that the method in this study has high precision and strong robustness in target tracking under the complicated conditions.]]>Morris, B.T. and M.M. Trivedi,2010255062Doucet, A., S. Godsill and C. Andrieu,200010197208Li, Y.Z., Z.Y. Lu, Q.X. Gao and J. Li,201032411415Luo, J. and Z.H. Wei,20091413611368Julier, S.J. and J.K. Uhlmann,200492401422Qu, Y.W., E.H. Zhang and J.Y. Yang,2010911521158Yu, M.Z., Z.X. Liu, J. Luo, L. Zheng-Xi, L. Jian and Y. Dan,201122766770Gong, J.L., X. He, Z.H. Wei and J.M. Guo,201220414421Liu, S.J., Z.M. Peng, S.B. Zhao and Q.H. Zhang,2010371623Wu, C. and C. Han,201021175179Shi, Y. and C.Z. Han,201137755759Yang, Y.X. and W.G. Gao,200680177183Gu, X., H.T. Wang, L.F. Wang, Y. Wang, R.B. Chen and C.H. Pan,201137550559Wang, X. and Z.M. Tang,201053280287Brasnett, P., L. Mihaylova, D. Bull and N. Canagarajaha,20072512171227Kim, B.G. and D.J. Park,20042517311742