基于规则面向对象的遥感影像分类方法在信息提取中的应用  

Application of rule-oriented object-based remote sensing image classification method in information extraction

作　　者：杨维超[1] YANG Wei-chao(Anhui Institute of Geophysical and Geochemical Exploration Technology,Hefei 230000,China)

机构地区：[1]安徽省地球物理地球化学勘查技术院,安徽合肥230000

出　　处：《世界有色金属》2020年第7期281-283,共3页World Nonferrous Metals

摘　　要：高空间分辨率遥感数据是遥感技术发展的一个重要方向.高分辨率遥感影像的应用也愈来愈广泛,包括建筑、矿山开采、国土、水利、电力等.但由于其数据量大、信息量丰富、光谱信息复杂,它的处理也成了一个难点.传统的基于像元的遥感分类技术对于高分辨率遥感影像的处理效果不够理想.本文以GEOEYE-1卫星影像为研究对象,采用基于规则的面向对象分类方法某区域用地信息进行提取.在envi的FX模块下,通过反复实验,确定各个区域最优分割与合并尺度,对影像进行多层次分割,通过检测对象的光谱、纹理、形状特征建立提取规则,对用地信息进行提取.研究表明,基于规则的面向对象分类方法分类精度为89.1%,kappa系数0.873,分类精度较高,可为信息提取提供参考.High spatial resolution remote sensing data is an important direction for the development of remote sensing technology.The application of high-resolution remote sensing images is becoming more and more extensive,including construction,mining,land,water conservancy,electricity,etc.However,due to its large amount of data,rich information,and complicated spectral information,its processing has also become a difficult point.The traditional pixel-based remote sensing classification technology is not ideal for processing high-resolution remote sensing images.This article takes GEOEYE-1 satellite imagery as the research object,and uses rule-based object-oriented classification method to extract land use information of a certain area.Under the FX module of envi,through repeated experiments,the optimal segmentation and merging scales of each region are determined,the image is divided into multiple levels,and the extraction rules are established by detecting the spectrum,texture,and shape features of the object to extract the land use information.The research shows that the rule-based object-oriented classification method has a classification accuracy of 89.1%,a kappa coefficient of 0.873,and a high classification accuracy,which can provide a reference for information extraction.

关 键 词：面向对象 影像分割与合并 规则 

分 类 号：TP751.1[自动化与计算机技术—检测技术与自动化装置]