论文部分内容阅读
遥感技术可快速准确地提取滑坡、崩塌、泥石流、堰塞湖等次生灾害信息。面向对象影像分类技术主要通过多尺度分割,并根据对象的特征进行模糊分类,可以很好的利用影像上下文信息,提高分类精度,与传统方法相比优势明显。利用汶川地震的震前、震后的福卫二号等数据进行研究,使用面向对象方法对北川附近灾情进行了变化检测分析;并结合DEM数据对唐家山堰塞湖璇坪乡区域震后初期的水体面积和蓄水增量进行了分析。在此基础上,总结了提取地震次生灾害的关键技术和工作流程。实验结果表明,应用面向对象的变化检测方法提取地震次生灾害可以取得很好的结果。
Remote sensing technology can quickly and accurately extract secondary disaster information such as landslide, collapse, debris flow and dammed lake. Object-oriented image classification technology is mainly based on multi-scale segmentation and fuzzy classification based on the characteristics of the object, which can make good use of the image context information to improve the classification accuracy, which has obvious advantages over the traditional methods. Using the data of Pre-earthquake and post-earthquake Fu-wei-2 data of the Wenchuan earthquake, the object-oriented method was used to analyze and analyze the disasters near Beichuan. Based on the DEM data, The water area and storage increment were analyzed. On this basis, the key technologies and workflow of extracting secondary disasters from earthquakes are summarized. Experimental results show that the application of object-oriented change detection method to extract secondary earthquakes can achieve good results.