基于混合PSO的高斯混合模型地形分类

Gaussian Mixture Model Terrain Classification Based on Hybrid PSO

提出了一种基于改进的混合粒子群优化(particle swarm optimization,PSO)算法的高斯混合模型地形分类方法。高斯混合模型的求解通常是使用期望最大化算法(expectation maximization,EM),然而EM算法易陷入局部最优,收敛速度不稳定且对初值敏感。因此引入混合PSO算法,并对其进行了一系列改进。实验结果表明:改进后的算法较其它优化算法提高了全局搜索能力和收敛速度,利用该算法求解高斯混合模型可以提高参数估计的精度,并且在户外场景图像的地形分类实验中所提出的地形分类方法也表现优良。

Gaussian mixture model terrain classification based on improved laybrid particle swarm optimization （PSO） algorithm was presented. The expectation-maximization （EM） method is a popular method to solve the Gaussian mix- ture model,but it is a local optimization method with instable convergence rate and initial value sensitivity. Therefore the hybrid PSO algorithm was introduced, and a series of improvement was conducted. Experimental results show that the improved algorithm can greatly improve the global convergence ability and enhance the rate of convergence. Using the improved algorithm to solve Gaussian mixture model can improve the accuracy of parameter estimation, and the pro- posed terrain classification method also has excellent performance in the terrain classification experiment of outdoor scene image.