OpenStreetMap辅助下的高分辨率光学影像道路损毁提取  

Damaged Road Extraction from High Resolution Optical Image Based on OpenStreetMap

作　　者：徐丰 董亮[2] 蔡肖芋 

机构地区：[1]民政部国家减灾中心,北京100124 [2]山西工程职业技术学院建筑工程系,山西太原030009 [3]武汉大学测绘遥感信息工程国家重点实验室,湖北武汉430079

出　　处：《地理空间信息》2016年第12期4-6,共3页Geospatial Information

基　　金：高分辨率对地观测系统重大专项技术研究与开发项目(03-Y20A10-9001-15/16):高分三号针对陆地目标的 SAR 图像质量控制研究技改项目

摘　　要：提出了一种以高分辨率遥感影像和OpenStreetMap道路矢量数据作为数据源的道路损毁提取方法。在OpenStreetMap道路矢量数据的辅助下,利用道路在高分辨率影像上的特征,结合学习-检测的方法对损毁区域进行检测,提取疑似损毁路段,基于道路信息对损毁路段进行验证,剔除虚警。实验表明,该方法能够快速、准确地对道路损毁信息进行提取。Because of the disasters occurrence, the characteristics of the roads have changed, so we have brought many difficulties to the extraction. Taking high resolution remote sensing images and Open Street Map as data sources, we proposed an automatic method for extracting damaged roads. Based on OpenS treetM ap road vector data, according to the characteristics of the roads in high resolution images, we used learning-detection method to extract the fracture zone of the road. Combining evidences and the observation of road fracture zone, we judged whether the fracture zone was damage area of the road. Experiment result indicates that the proposed method can extract road damage information quickly and accurately.

关 键 词：高分辨率影像 OpenStreetMap 学习-检测 道路损毁提取 

分 类 号：P237[天文地球—摄影测量与遥感]