嵌入卡尔曼预测器的粒子滤波目标跟踪算法.pdfVIP

27 2 Vo.l 27No. 2 20102 pplicationResearch ofComputers Feb. 2010 * , ( , 315211) : , , , , , , , : ; ; ; : TP309 : : 100123695(2010)0220468204 doi: 10. 3969/j. issn. 100123695. 2010. 02. 017 Particle filter and embedded inKaml an predictor for target tracking W NG Yue2ling, W NG Rang2ding (CKC Software Labora tory, N ingbo Un iversity, N ingbo Zhejiang 315211, China ) Ab stra ct: The standard colorparticle filter spreads each particlewith random walkmethod for target tracking. The traditional colorhistogram isused that can not reflect the characteristicsofthe target space. Thispaperproposed anovel color2based par2 ticle filter target tracking algorithms. Thesecond2order colorhistogramwas applied to get theobservation ofparticleprobability density function. In addition, used theKaml an filter to determ ine the spread ofparticle dynam icmodel of the uncertainty in the drift of the state. Thedistribution ofparticleswasmore accurately close to the probability distribution of the target, thus the use efficiency ofparticleswas greatly mi proved. Computer smi ulation results demonstrate that the proposed algorithm is more robust as compare to the traditional color2based particle filter tracking algorithm. K ey w or ds: colorhistogram; particle filter(PF); Kaml an predictor; target tracking Kalman [ 1] , , , ; (), ; ,