基于粒子群优化算法的航空影像纹理分类  

Aerial Image Texture Classification Based on Particle Swarm Optimization

作　　者：李林宜[1] 李德仁[2] 

机构地区：[1]武汉大学遥感信息工程学院,武汉430079 [2]武汉大学测绘遥感信息工程国家重点实验室,武汉430079

出　　处：《计算机工程》2008年第6期196-198,共3页Computer Engineering

基　　金：国家自然科学专项基金资助项目(40523005)

摘　　要：粒子群优化算法是基于群智能的随机全局优化技术,将它引入航空影像纹理分类,在提取纹理样本小波、分形等特征的基础上,提出了针对分类问题的粒子表达方法和群体寻优策略,实现了基于粒子群算法的纹理分类。将其与基于遗传算法的纹理分类法作比较,结果表明粒子群优化算法具有较好的寻优性能,基于该算法的纹理分类法分类精度较高且计算时间较少。Particle Swarm Optimization（PSO） is a stochastic global optimization technique based on swarm intelligence. This paper introduces aerial image texture classification. The particles in the swarm are constructed and swarm search strategies are proposed in terms of the needs of classit^cation application after many texture features such as wavelet features and factual features extracted from texture samples. The classification method as PSO is implemented. Compared with the texture classification method based on Genetic Algorithms（GA）, PSO has better search ability and the classification method based on it has higher classification accuracy and needs less computation time in the experiment.

关 键 词：粒子群优化算法 纹理分类 航空影像 特征提取 遗传算法 

分 类 号：TP391.41[自动化与计算机技术—计算机应用技术]