复杂背景下基于改进粒子滤波的红外人体跟踪

Infrared Human Tracking Based on Improved Particle Filter Under Complex Background

为解决复杂背景下红外图像序列中的人体跟踪问题,提出了一种改进的粒子滤波跟踪方法。根据红外图像中人体目标的特点,首先建立人体的灰度直方图来提取其灰度特征,同时采用一种新的基于帧间差分和灰度概率分布图的方法提取其运动特征。然后将上述两种特征融合到粒子滤波框架中,用于粒子权值的计算,最终实现红外序列中人体的稳健跟踪。实验结果表明,和传统粒子滤波算法相比,该方法大大提高了复杂背景下红外人体跟踪的准确性和有效性,跟踪结果令人满意。

An improved particle filter tracking algorithm was proposed to solve the problem of human tracking in infrared image sequences under complex background.According to the characters of the human in the infrared images,the algorithm firstly constructed the gray histogram of the human to extract the gray feature,and at the same time utilized a new method based on the inter-frame difference and gray probability distribution image to get the motion feature.Then,the above-mentioned two features were fused into the particle filter frame to calculate the particle weights.Finally the robust tracking of human in infrared image sequences was achieved.The experimental results show that compared with the traditional particle filter algorithm,the presented method greatly improves the accuracy and effectiveness of the infrared human tracking under complex background,and the tracking results are satisfactory.