基于模糊聚类和粒子滤波目标跟踪算法研究

摘要

A hybrid algorithm based on the fuzzy clustering and the particle filter is adopted focusing on the target tracking in the nonlinear model, and the good effect is made better than the Kalman filter. Firstly, the data collected is classified by the fuzzy c-means clustering algorithm. Secondly, the target position is estimated through the particle filter algorithm. The simulation results show that the target tracking effect through the hybrid method is better than the Kalman filtering algorithm in the nonlinear model and the tracking error is reduced.%针对非线性的目标跟踪采用了基于模糊聚类和粒子滤波的混合跟踪算法,取得了优于卡尔曼滤波跟踪的良好效果.首先利用模糊C均值聚类算法对采集的数据进行基于目标的隶属度的分类,然后利用粒子滤波算法对目标进行位置估计.仿真结果表明:非线性视频跟踪中混合算法对目标的跟踪效果要好于卡尔曼滤波算法,降低了跟踪误差.