利用结构面对象的高分遥感影像建筑物变化检测  

High-resolution remote sensing image building change detection using structural objects

作　　者：刘海飞[1,2] 杨敏华[2] 王晶[2] 徐刚[2,3] 陈杰 LIU Haifei;YANG Minhua;WANG Jing;XU Gang;CHEN Jie(The Third Surveying and Mapping Institute of Hunan Province,Changsha 410007,China;Central South University,Changsha 410083,China;Zhejiang Institute of Security Technology,Wenzhou 325016,China)

机构地区：[1]湖南省第三测绘院,湖南长沙410007 [2]中南大学,湖南长沙410083 [3]浙江安防职业技术学院,浙江温州325016

出　　处：《测绘通报》2021年第6期16-20,共5页Bulletin of Surveying and Mapping

摘　　要：根据建筑物具有丰富的角点和规则的外观形状的特点,本文提出了依据线段信息生成建筑物变化检测基本单元——结构面对象的建筑物变化检测方法。从高分遥感影像中提取出建筑物的边缘线及屋顶外形线条,生成结构面对象并提取其在两时相影像中相应的特征信息,通过监督分类获取变化建筑区域。试验结果表明,利用结构面对象进行建筑物变化检测能较完整地覆盖变化建筑物表面且对象边缘检测效果较好。According to the building with rich corner points and regular appearance shape characteristics,a method of building change detection based on line segment information to generate the basic unit of building change detection—the structural surface object is proposed.The edge lines and roof contour lines of buildings are extracted from high-resolution remote sensing images,then the structural surface objects are generated and the corresponding feature information is extracted from the two-phase images.And then,we obtain the changed building area through supervised classification.The experimental results show that the use of structural surface objects for building change detection can better cover the surface of the changed building and the object edge detection effect is better.

关 键 词：建筑物变化检测 结构面对象 线特征 结构特征 超像素 

分 类 号：P237[天文地球—摄影测量与遥感]