Damaged building detection based on GF-1 satellite remote sensing image: A case study for Nepal MS8.1 earthquake
-
摘要: 在高分辨率遥感图像中, 不同震害损毁程度的建筑物呈现不同的图像特征, 鉴于此本文提出一种利用遥感图像多特征分析建筑物损毁程度的检测方法. 以2015年尼泊尔MS8.1地震为例, 结合震后高分一号卫星全色遥感图像和城市道路矢量数据提供的街区信息, 以建筑物街区为单元进行图像纹理提取和局部空间统计等多类别图像特征参数分析, 并构建多特征分类模型, 将震后建筑物街区划分为基本完好、 部分损毁和严重损毁等3个类别. 试验结果表明, 本文提取的参数能够有效地表征损毁建筑物的图像特征, 而且建筑物震害损毁检测精度较高. 该方法可用于建筑物震害损毁信息的快速提取, 为震后应急救援提供指导; 同时还可为我国自主研发高分卫星遥感数据在地震灾害信息提取中的应用提供技术参考与方法借鉴.Abstract: The damaged buildings caused by an earthquake present different image features from that of the intact buildings, therefore the building damage information could be distinguished from a variety of target features. From this point of view, this paper proposed an approach for building damage detection by utilizing various features of the building blocks. Taking the GF-1 satellite image of the Nepal MS8.1 earthquake occurred in 2015 as an example, this paper utilized blocks information provided by the GIS data, and classified the building blocks in the studied area into three categories of the intact, partly damaged and seriously destroyed, on the basis of the quantitative analysis results about texture features of remote sensing image and local spatial statistics of the building blocks. The test results demonstrated that the indicative key parameters extracted in this paper could effectively demonstrate the image characteristics of the damaged building, so that we can effectively conduct the classification and detection of building damage information caused by earthquakes with the proposed detection method in this paper. Also, it could provide guidance for earthquake emergency rescue, and it provides technical experiences and references for the building damage detection using the GF-1 data with independent intellectual property rights in our country.
-
Keywords:
- GF-1 satellite /
- earthquake /
- building damage detection /
- remote sensing image
-
-
表 1 建筑物损毁检测试验精度评价
Table 1 Accuracy assessment of building damage detection
检测结果 参考结果中各类损毁目标的数量 严重损毁 部分损毁 基本完好 合计 严重损毁 5 2 2 9 部分损毁 2 28 1 31 基本完好 1 5 4 10 合计 8 35 7 50 生产者精度 62.50% 80.00% 57.14% 用户精度 55.56% 90.32% 40.00% 总体精度 74.00% -
陈文凯. 2007. 面向震害评估的遥感应用技术研究[D]. 兰州: 中国地震局兰州地震研究所: 2-8. Chen W K. Research of Riote Sensing Application Technology Based on Earthquake Disaster Assessment[D]. Lanzhou: Lanzhou Institute of Seismology, China Earthquake Administration: 2-8 (in Chinese).
曾涛. 2010. 汶川地震重灾区多源影像处理及震害信息提取方法研究[D]. 成都: 成都理工大学: 12-13. Zeng T. 2010. Methods of Multi-Sensors RS Image Processing and Earthquake Damage Information Extraction for the Wenchuan Earthquake Hit-Area[D]. Chengdu: Chengdu University of Technology: 12-13 (in Chinese).
中国地震台网中心. 2015. 尼泊尔8.1级地震[EB/OL]. [2015-04-25]. http://news.ceic.ac.cn/CC20150425141126.html. China Seismic Networks Center. 2015. Nepal MS8.1 earthquake[EB/OL]. [2015-04-25]. http://news.ceic.ac.cn/CC20150425141126.html (in Chinese).
Nepal Disaster Risk Reduction Portal. 2015. Nepal earthquake 2015: Disaster relief and recovery information platform (NDRRIP)[EB/OL]. [2015-04-25]. http://drrportal.gov.np/.
Samadzadegan F, Rastiveisi H. 2008. Automatic detection and classification of damaged buildings, using high resolution satellite imagery and vector data[C]//The International Archives of the Photogrammetry, Riote Sensing and Spatial Information Sciences, Vol.37. Beijing: XXIst ISPRS Congress: 415-420.
Sumer E, Turker M. 2005. Building damage detection from post-earthquake aerial imagery using building grey-value and gradient orientation analyses[C]//Recent Advances in Space Technologies, 2005, Proceedings of 2nd International Conference. Istanbul: IEEE: 577-582.
United Nations Institute for Training and Research. 2015. Damage assessment of Bhaktapur, Kathmandu Valley, Nepal[EB/OL]. [2015-04-30]. http://www.unitar.org/unosat/node/44/2205?utm_source=unosat-unitar&utm_medium=rss&utm_campaign=maps.