摘要
土地利用信息是进行土地规划和管理的重要数据,有着重要的经济价值。采用计算机仿真技术对遥感影像进行自动分类是一种获取土地利用数据十分有效的手段。然而遥感影像的不确定、不一致现象易导致过度拟合,增加了分类难度。提出了一种新的基于粗集的决策树用于遥感影像分类。经试验表明该分类方法较CART树、ID3树等算法在分类精度、防止过度拟合方面均有所提高。
Land cover information has been identified as the crucial data for land planning and management,which has important economic value.In order to obtain land cover information,utilizing computer simulation technology to automatically classify the remote sensing images is a very effective measure.However,remote sensing image's uncertainty,inconsistency may lead to the over-fitting phenomenon and increase the difficulty of classification.This paper proposed a new method of remote sensing decision tree based on rough set.Simulation results show that this method can reduce over-fitting phenomenon and improve the classification accuracy than CART tree and ID3 tree method.
出处
《长春工程学院学报(自然科学版)》
2010年第4期95-97,共3页
Journal of Changchun Institute of Technology:Natural Sciences Edition
基金
吉林省科技厅青年基金项目(20090120)和(20100190)
关键词
粗集
决策树
遥感影像
监督分类
rough set
decision tree
remote sensing
supervised classification