基于粒子滤波的灰度目标跟踪算法

A Gray Object Tracking Algorithm Based on Particle Filter

粒子滤波主要利用粒子集来表示概率,可以用在任何形式的状态空间模型上.提出了一种基于粒子滤波的灰度图像目标跟踪方法,粒子滤波适合各种形式状态空间模型.算法目标特征采用了灰度直方图、灰度梯度直方图对灰度图像序列进行跟踪.粒子滤波跟踪算法有状态转移和状态观测两大重要模型.利用高权值的粒子替代低权值粒子这样的粒子重采样来保证粒子集的健壮性,得到目标最终位置.利用Matlab进行仿真证明了本文算法的有效性和稳健性.

With the development of science and technology,tracking Technology is widely used in the daily life. This paper proposed a target tracking method of gray level image based on particle filter which is fit for all kinds of form state space model. The algorithm present in this article makes use of the gray information and the gradient characteristics of the target to track the object in gray level images. A state transfer and observe model of the system is constructed in this method. In order to guarantee the robustness of the particle set,particles of high weights are used to instead of the low ones to sample the particles and finally obtain the location of the target. The effectiveness and robustness of this algorithm is simulated by Matlab.