蜂群智能算法的遥感影像分类方法  

An Innovative Method to Classify Remote Sensing Images Using Artificial Bee Colony Algorithm

作　　者：曹敏[1] 史照良[2,3] 阳建逸[1] 

机构地区：[1]南京师范大学虚拟地理环境教育部重点实验室,江苏南京210023 [2]国土环境与灾害监测国家测绘地理信息局重点实验室,江苏徐州221116 [3]江苏省地理信息技术重点实验室,江苏南京210093

出　　处：《测绘学报》2013年第5期745-751,共7页Acta Geodaetica et Cartographica Sinica

基　　金：国家自然科学基金(41101349);江苏省测绘科研资助项目(JSCHKY201304);国土环境与灾害监测国家测绘地理信息局重点实验室开放基金(LEDM2012B07);江苏省高校自然科学研究项目(13KJB420003);江苏高校优势学科建设工程

摘　　要：结合仿生智能计算方法,提出一种基于人工蜂群算法的智能式遥感分类方法。采用模仿蜜蜂采蜜行为的蜂群智能优化算法,自动搜索遥感影像各波段的最优分割点,定义各波段最优分割点和地物类别节点的连线为蜂群搜索路径,构造以IF-THEN形式表达的遥感分类规则。该智能式遥感分类方法不需使用数学公式,能更清晰地表达地物分类中的复杂关系。将该方法应用于长江口北岸ALOS影像分类中,取得了较好的分类结果,并与See5.0决策树方法进行了对比分析。研究表明,蜂群智能遥感分类的总体精度和Kappa系数均比See5.0决策树方法更高,该智能遥感分类方法具有更好的分类效果。By the combination of the bio-inspired computing methods, an intelligent method to classify remote-sens- ing images using artificial beecolony（ABC） algorithm is proposed. This ABC-based classification method searches for the optimal split points on each band of remote-sensing image automatically, where the optimal solution of split points emerge from intelligent behaviour of honey bee swarms. The method defines bees search path as the connec- tion between optimal split points and feature category node, and constructs classification rules in the form of ＇IF- THEN~. ThJssmart approach can express the complex classification more clearly without mathematical formulas. By taking an example of ALOS image in the north shore of the Yangtze River estuary, the proposed classification method based on ABC algorithm were implemented and tested against SeeS. 0 decision-tree method. The overall classification accuracy and Kappa coefficient of ABC intelligent method are higher than the See5. 0 decision-tree one. It is shown that theABC intelligent method has better classification results.

关 键 词：人工蜂群算法 仿生智能计算 遥感影像 分类 

分 类 号：P237[天文地球—摄影测量与遥感]