基于假设检验的自适应粒子滤波红外目标跟踪

Infrared Target Tracking for Auto-adaptive Particle Filtering Based on Hypothesis Test

在实时性要求较高的非线性非高斯环境中,粒子滤波中的粒子数选取将直接影响红外目标跟踪效果。为此,提出一种基于假设检验的自适应粒子滤波算法。通过假设检验问题中样本容量的选取确定粒子数,解决因粒子数过大造成的时间损耗。实验结果表明,该算法在保证目标跟踪准确度的同时可减少跟踪延时,具有较好的实时跟踪效果。

Particle filtering is taken as the main solution for solved infrared target tracking problems in nonlinear/non-Gaussian system.The selection of number of particle set directly influences on the tracking effect on which particle filtering dose in high real-time system.Based on the problem,a novel algorithm based on hypothesis test in auto-adaptive particle filtering is proposed for target tracking in infrared imagery real-time system.According to the hypothesis test problems,the number of particle set is dynamic got,which works out the problem of high time consume that larger number of particle set is taken.Experimental results performed on several infrared image sequences show the robustness and better real-time performance of the proposed algorithm.