中文摘要：

利用DigitalGlobal提供的遥感影像数据,提取了尼泊尔震区某一建筑区域地震前后的影像。采用监督分类、非监督分类和面向对象的分类方法对震前震后的遥感图像进行了分类,对3种分类结果进行了对比。结果表明,面向对象的分类方法比其他两种方法有更好的分类效果。通过与ArcGIS的人工矢量化图进行对比,发现在进行建筑物数量提取中,面向对象的分类方法能较准确地提取建筑物数量,为快速获取地震灾情、进行烈度评定等提供依据,对政府相关部门的应急救援与指挥决策有一定的指导作用。

英文摘要：

Using remote sensing image data provided by Digital Global, this paper extracted the image of a building area in the Nepal earthquake region before and after the earthquake . BY three methods of supervised classifica- tion, unsupervised classification and object - oriented classification, the remote sensing images before and after the earthquake were classified, and the classification results of the three methods were compared. Results show that, the object -oriented classification method had good classification effect than other methods. In addition, the object - oriented classification method can extract the building number more accurately than the artificial vectorization fig- ure of ArcGIS. Therefore, the method can provides basis and support for works such as earthquake disaster situation acquisition, intensity evaluation and so on, and play a certain guiding role in emergency rescue and command deci- sion- making of relevant government departments.