针对红外图像消噪的特性，提出粒子重采样滤波算法。首先通过Chapman-Kolmogorov方程对粒子群系统空间状态的概率密度函数预测状态；然后重采样去除小权值粒子，保留复制权值较大的粒子，且大权值粒子多次采样；接着有效粒子数阈值防止粒子退化，划分粒子权值为大、中、小三类，中权值粒子保留，大、小权值粒子合并产生新粒子，通过Thompson-Taylor算法随机挑选新粒子重采样；最后消噪模型采用两种噪声迭加成的混合双模噪声模型，给出了算法流程。仿真结果表明，本文算法在有效保留图像重要信息的同时对噪声的抑制效果更为理想。%According to the characteristics of infrared image denoising, resample filter algorithm is presented. Firstly, the probability density function of Chapman-Kolmogorov equation is used to predict the state of the particle swarm system space, then small particle is removed in accordance with resample weights, those particles with larger weight is retained, and are sampled many times; then the effective particle number threshold is set to prevent particle degradation in particle weight, which is divided into large, medium, and small ones. Medium weight particle being kept, big and small weight particles are combined to generate new particles. Through the Thompson-Taylor algorithm new particles are randomly selected to be resampled. Finally denoising model using two kinds of noise hybrid Bimodal Noise Model and the algorithm flow is given. The simulation results show that the algorithm proposed in this paper not only denoise image effectively but also retain important information at the same time.
展开▼
