Object-oriented information extraction and evaluation of seismic damage of buildings using very high spatial resolution imagery

摘要

Earthquake is one of the most serious natural disasters in the world. With the continuous improvement of the spatial resolution of remote sensing image, it contains more abundant information, which can better reflect the features of geometric structure and texture information. The operation of the image is not only rely on the image element, but also use the object to carry on the operation. This paper proposes a fast building damage information extraction and evaluation of technology process which can quickly extract and evaluate the disaster area. Using the method of object oriented extracts the buildings of Yushu disaster area based on the pre-earthquake and post-earthquake QuickBird image. It has a high accuracy of 88.53% and 90.21%, respectively, and extracts the regional change information, whose change area is 15923.52m2 accounted for 68.16% of the entire study area that assesses for moderate to severe disaster areas. This rapid seismic damage information extraction is a kind of practical and effective method, which is worth popularizing widely.