| 基于蛙跳算法的离散粒子群优化端元提取 |
| |
| 引用本文： | 吴国伟,赵艳玲,王龙,倪巍,许立江,冉艳艳.基于蛙跳算法的离散粒子群优化端元提取[J].中国图象图形学报,2015,20(5):724-732. |
| |
| 作者姓名： | 吴国伟 赵艳玲 王龙 倪巍 许立江 冉艳艳 |
| |
| 作者单位： | 中国矿业大学(北京), 地球科学与测绘工程学院, 土地复垦与生态重建研究所, 北京 100083;中国矿业大学(北京), 地球科学与测绘工程学院, 土地复垦与生态重建研究所, 北京 100083;中国矿业大学(北京), 地球科学与测绘工程学院, 土地复垦与生态重建研究所, 北京 100083;中国矿业大学(北京), 地球科学与测绘工程学院, 土地复垦与生态重建研究所, 北京 100083;中国矿业大学(北京), 地球科学与测绘工程学院, 土地复垦与生态重建研究所, 北京 100083;中国矿业大学(北京), 地球科学与测绘工程学院, 土地复垦与生态重建研究所, 北京 100083 |
| |
| 基金项目： | 教育部新世纪优秀人才支持计划(NCET-12-0964) |
| |
| 摘 要： | 目的 针对离散粒子群优化(D-PSO)端元提取算法易“早熟”,易陷入局部最优解等问题,引入蛙跳算法,提出了基于蛙跳算法的离散粒子群优化(SFLA-DPSO)端元提取算法.方法 该算法把粒子群分成若干族群,先在每个族群内进行深度寻优,然后在族群间完成信息交流,实现了SFLA算法的全局性、并行性与D-PSO算法的快速收敛性相结合,进而避免粒子陷入局部最优解.分别用SFLA-DPSO、D-PSO和SMACC对云南普朗地区Hperion高光谱影像提取端元;同时,在Hperion和AVIRIS高光谱影像的可行解搜索空间内,分别用SFLA-DPSO、D-PSO和N-FINDR提取端元,借助统计学理论分析计算两种算法在不同迭代次数下达到全局收敛的概率.结果 当达到一定迭代次数后,SFLA-DPSO出现全局收敛的概率基本达到100%,而D-PSO却仅在65%左右,因此SFLA-DPSO算法具有较高的可信度.结论 从而认为SFLA-DPSO克服局部收敛的能力更强,表现出良好的稳定性.
|
| 关 键 词： | 离散粒子群优化 蛙跳算法 端元提取 高光谱遥感 局部收敛 |
| 收稿时间： | 2014/9/16 0:00:00 |
| 修稿时间： | 2014/12/26 0:00:00 |
Discrete particle swarm optimization endmember extraction based on shuffled frog leaping algorithm |
| |
| Wu Guowei,Zhao Yanling,Wang Long,Ni Wei,Xu Lijiang and Ran Yanyan.Discrete particle swarm optimization endmember extraction based on shuffled frog leaping algorithm[J].Journal of Image and Graphics,2015,20(5):724-732. |
| |
| Authors: | Wu Guowei Zhao Yanling Wang Long Ni Wei Xu Lijiang and Ran Yanyan |
| |
| Affiliation: | Institute of Land Reclamation and Ecological Restoration, College of Geoscience and Surveying Engineering, China University of Mining & Technology(Beijing), Beijing 100083, China;Institute of Land Reclamation and Ecological Restoration, College of Geoscience and Surveying Engineering, China University of Mining & Technology(Beijing), Beijing 100083, China;Institute of Land Reclamation and Ecological Restoration, College of Geoscience and Surveying Engineering, China University of Mining & Technology(Beijing), Beijing 100083, China;Institute of Land Reclamation and Ecological Restoration, College of Geoscience and Surveying Engineering, China University of Mining & Technology(Beijing), Beijing 100083, China;Institute of Land Reclamation and Ecological Restoration, College of Geoscience and Surveying Engineering, China University of Mining & Technology(Beijing), Beijing 100083, China;Institute of Land Reclamation and Ecological Restoration, College of Geoscience and Surveying Engineering, China University of Mining & Technology(Beijing), Beijing 100083, China |
| |
| Abstract: | |
| |
| Keywords: | discrete particle swarm optimization huffled frog leaping algorithm endmember extraction hyperspectral remote sensing local convergence |
|
| 点击此处可从《中国图象图形学报》浏览原始摘要信息 |
|
点击此处可从《中国图象图形学报》下载免费的PDF全文 |
|
