Extraction Method for Earthquake-Collapsed Building Information Based on High-Resolution Remote Sensing

35th International Symposium on Remote Sensing of Environment

地球科学;生态环境科学

Chen, Peng^1 ; Wu, Jian^1,4 ; Liu, Yaolin^2 ; Wang, Jing^3

State Research Centre of Island Planning and Protecting, Third Institute of Oceanography State Oceanic Administration, 178 Daxue Road, Xiamen, China^1

School of Resource and Environmental Science, Wuhan University, 129 Ruoyu Street, Wuhan, China^2

Key Laboratory of Land Use, China Land Surveying and Planning Institute, 37 Guanyingyuanxiqu, Beijing, China^3

No. 178, Daxue Road, Xiamen City, Fujian Province, China^4

关键词: Earthquake disaster informations;    High resolution remote sensing;    Object oriented method;    Object oriented theory;    Object-oriented image analysis;    Precision assessment;    Traditional techniques;    Visual interpretation;   

Others  :  https://iopscience.iop.org/article/10.1088/1755-1315/17/1/012096/pdf
DOI  :  10.1088/1755-1315/17/1/012096

学科分类：环境科学（综合）

来源: IOP

【 摘 要 】

At present, the extraction of earthquake disaster information from remote sensing data relies on visual interpretation. However, this technique cannot effectively and quickly obtain precise and efficient information for earthquake relief and emergency management. Collapsed buildings in the town of Zipingpu after the Wenchuan earthquake were used as a case study to validate two kinds of rapid extraction methods for earthquake-collapsed building information based on pixel-oriented and object-oriented theories. The pixel-oriented method is based on multi-layer regional segments that embody the core layers and segments of the object-oriented method. The key idea is to mask layer by layer all image information, including that on the collapsed buildings. Compared with traditional techniques, the pixel-oriented method is innovative because it allows considerably rapid computer processing. As for the object-oriented method, a multi-scale segment algorithm was applied to build a three-layer hierarchy. By analyzing the spectrum, texture, shape, location, and context of individual object classes in different layers, the fuzzy determined rule system was established for the extraction of earthquake-collapsed building information. We compared the two sets of results using three variables: precision assessment, visual effect, and principle. Both methods can extract earthquake-collapsed building information quickly and accurately. The object-oriented method successfully overcomes the pepper salt noise caused by the spectral diversity of high-resolution remote sensing data and solves the problem of same object, different spectrums and that of same spectrum, different objects. With an overall accuracy of 90.38%, the method achieves more scientific and accurate results compared with the pixel-oriented method (76.84%). The object-oriented image analysis method can be extensively applied in the extraction of earthquake disaster information based on high-resolution remote sensing.