计算机辅助设计与图形学学报
計算機輔助設計與圖形學學報
계산궤보조설계여도형학학보
JOURNAL OF COMPUTER-AIDED DESIGN & COMPUTER GRAPHICS
2009年
12期
1774-1784
,共11页
袁广林%薛模根%谢恺%姚翎
袁廣林%薛模根%謝愷%姚翎
원엄림%설모근%사개%요령
目标跟踪%核函数%粒子滤波%特征融合
目標跟蹤%覈函數%粒子濾波%特徵融閤
목표근종%핵함수%입자려파%특정융합
target tracking%kernel function%particle filter%features fusion
经典粒子滤波及其改进算法在观测模型与真实情况存在偏差时会导致滤波发散,针对这一问题,提出一种核函数粒子滤波算法.该算法根据目标状态与粒子状态之间的距离,利用核函数产生权值对粒子进行二次加权,根据粒子的二次加权结果进行粒子重采样;以改进的粒子滤波算法为框架,提出了一种自适应多特征融合目标跟踪方法,利用相似性度量动态地评价特征对目标与背景的区分能力,并自适应地计算特征融合权重,以适应目标跟踪过程中目标与背景的变化,提高目标跟踪的鲁棒性.实验结果表明,文中提出的目标跟踪方法比经典粒子滤波目标跟踪方法具有更强的抗干扰性能和较高的跟踪精度.
經典粒子濾波及其改進算法在觀測模型與真實情況存在偏差時會導緻濾波髮散,針對這一問題,提齣一種覈函數粒子濾波算法.該算法根據目標狀態與粒子狀態之間的距離,利用覈函數產生權值對粒子進行二次加權,根據粒子的二次加權結果進行粒子重採樣;以改進的粒子濾波算法為框架,提齣瞭一種自適應多特徵融閤目標跟蹤方法,利用相似性度量動態地評價特徵對目標與揹景的區分能力,併自適應地計算特徵融閤權重,以適應目標跟蹤過程中目標與揹景的變化,提高目標跟蹤的魯棒性.實驗結果錶明,文中提齣的目標跟蹤方法比經典粒子濾波目標跟蹤方法具有更彊的抗榦擾性能和較高的跟蹤精度.
경전입자려파급기개진산법재관측모형여진실정황존재편차시회도치려파발산,침대저일문제,제출일충핵함수입자려파산법.해산법근거목표상태여입자상태지간적거리,이용핵함수산생권치대입자진행이차가권,근거입자적이차가권결과진행입자중채양;이개진적입자려파산법위광가,제출료일충자괄응다특정융합목표근종방법,이용상사성도량동태지평개특정대목표여배경적구분능력,병자괄응지계산특정융합권중,이괄응목표근종과정중목표여배경적변화,제고목표근종적로봉성.실험결과표명,문중제출적목표근종방법비경전입자려파목표근종방법구유경강적항간우성능화교고적근종정도.
The particle filtering has been extensively used for visual tracking due to its flexibility. However the conventional particle filtering and its improved variants usually diverge when the measurement model is not accurate enough. To address this problem, a kernel-based particle filter algorithm is proposed. The algorithm reweighs the particles by weights which are produced by kernel function with the distance between target state and particles, and the particles are resampled according to the resultant weights. With the above improved particle filter algorithm, an adaptive multiple features fusion target tracking method is proposed. The proposed tracking method dynamically assesses the discriminability of each feature with respect to foreground to background separability and adaptively computes the feature's fusion weight by some similarity measure. Experimental results show that the proposed tracking method is superior over the conventional particle filter based tracking methods.