文章摘要

文章信息/Info

Title:: Research on Spatialization of Buildings in Longnan City,Gansu Province Based on Miniature Unmanned Aerial Vehicle and High Resolution Satellite Images

作者:: 陈晋¹; 习聪望²; 陈文凯^1; 张苏平¹; 周中红¹; (1.中国地震局兰州地震研究所,甘肃兰州 730000; 2.陕西省地震局,陕西西安 710068)

Author(s):: CHEN Jin¹; XI Congwang²; CHEN Wenkai¹; ZHANG Suping¹; ZHOU Zhonghong¹; (1. Lanzhou Earthquake Research Institute,China Earthquake Administration,Gansu 730000,Lanzhou,China)(2. Shaanxi Earthquake Agency,Xi’an 723000,Shaanxi,China)

Keywords:: unmanned aerial vehicle(UAV); high resolution satellite image; residential area; building type; spatialization

摘要:: 基于无人机、高分卫星影像资料,通过实地调研与遥感影像对比分析,建立基于无人机、高分卫星遥感影像获取建筑物的技术路线,并以甘肃省陇南市为研究区进行实例验证。研究结果表明:利用无人机航拍进行建筑物识别时,采用倾斜摄影和正射影像相结合的方式,建筑物识别效果较好,尤其是对屋顶相同或类似的不同结构建筑物的识别;基于遥感技术获取建筑物时不仅要建立区域建筑物遥感影像解译标志,还需要借助区域地理环境特征、建筑物排列、占地面积、建筑物阴影等因素进行辅助识别,才能获取较为可靠的结果; 陇南市建筑物类型主要有土木(含木构架)结构、砖木结构、砖混结构、框架结构4类,占比分别为19.25%、44.29%、31.32%、5.14%,建筑物遥感解译结果精度在-23.92%～25.28%; 基于无人机和卫星遥感影像获取居民地建筑物数据可以用于更新地震应急基础数据库,但存在一定的误差。

Abstract:: Based on miniature unmanned aerial vehicle(UAV),high resolution satellite images and comparison between field survey data and remote sensing image,we build a technical route to get building data through UAV and high resolution satellite images,which is validated by taking Longnan city as the study area. The research result shows that the combined use of unmanned aerial vehicle and tilt photography can lead to a better result for building recognition,especially for the different types of buildings with similar roofs. The establishment of building interpretation signs can lead to a more reliable result with the help of geographical environment characteristics,building arrangement,building area,building shadows and other factors. The building types of Longnan city are mainly civil(including wood framestructure,brick-wood structure,brick-concrete structure,frame structure,and the proportion of each type reaches 19.25%(civil structure),44.29%(brick-wood),31.32%(brick-concrete structure),5.14%(frame structure),respectively. The accuracy range of the interpretation varies from -23.92% to 25.28%. It is feasible to obtain the residential building data for the update of the earthquake emergency database based on UAV and the satellite remote sensing images though there is some errors.