基于粒子滤波的视频目标跟踪算法研究及改进

本文系统地研究了粒子滤波理论和基于粒子滤波的视频目标跟踪算法应用,总结归纳了标准粒子滤波算法存在的不足并提出相应的改进方法。针对目标的表示方法,加入了梯度直方图特征,并对重采样进行了改进。实验结果表明,改进后的跟踪算法,在算法的准确性、鲁棒性和实时性上都得到了较大的提高。

This paper systematically studies the particle filter theory and the application of video target tracking algorithm based on particle filter.It sums up the shortcomings of the standard particle filter algorithm and proposes corresponding improvement methods.For the representation of the target,the gradient histogram feature was added and the resampling was improved.Experimental results show that the improved tracking algorithm has been greatly improved in accuracy,robustness,and real-time performance of the algorithm.