一种基于图像熵的密度峰值聚类波段选择方法  

A Method for Density Peaks Clustering Band Selection Based on Image Entropy

作　　者：赵海士 路来君[1] 杨晨[1] ZHAO Haishi LU Laijun YANG Chen(College of Earth Sciences, Jilin University, Changchun 130061, China)

机构地区：[1]吉林大学地球科学学院,长春130061

出　　处：《吉林大学学报（理学版）》2017年第2期376-378,共3页Journal of Jilin University:Science Edition

基　　金：国家重点基础研究发展计划973项目基金(批准号:2015CB453000);国家自然科学基金(批准号:61572228;61272207;11403049);吉林省科技发展计划项目(批准号:20160101247JC;20140520070JH)

摘　　要：基于将图像熵引入到密度峰值聚类算法中以确定波段信息量,提出一种基于图像熵的密度峰值聚类波段选择方法.通过构建衡量波段重要性得分,解决了传统密度峰值聚类方法在波段选择时未考虑波段信息量的问题.实验结果表明,该方法的分类精度较传统密度峰值聚类方法平均提高2.12%.Based on the introduction of image entropy into the density peaks clustering algorithm to determine the information quantity of spectral band, we proposed a method for density peaks clustering band selection based on image entropy. By constructing the measuring score of bands significance, we solved the problem of the band selection in the traditional density peaks clustering method without considering the information quantity of the hands. The experimental results show that the classification accuracy of the proposed method increases by an average of 2.12% than that of the traditional density peaks clustering method.

关 键 词：波段选择 图像熵 高光谱遥感 密度峰值聚类 

分 类 号：TP751[自动化与计算机技术—检测技术与自动化装置]