“高分二号”卫星数据遥感滑坡灾害识别研究--以云南东川为例

摘要

标志着中国遥感卫星进入亚米级“高分时代”的“高分二号”（GF-2）卫星在太原成功发射，它将为地质灾害调查提供可靠的数据支持。文章以滑坡灾害频发的云南东川区为研究对象，首先对GF-2卫星采集的研究区数据进行正射校正、配准、融合和裁剪等预处理；然后采用面向对象的分类方法，通过影像分割、合并分块，并结合滑坡灾害的光谱、纹理和数字高程模型等多种信息特征，建立滑坡灾害识别规则，实现滑坡信息的遥感分类；最后对分类结果进行误分对象剔除，得到研究区滑坡信息。结果显示，利用 GF-2卫星数据可以较好地提取滑坡灾害信息，基本满足滑坡的灾害识别要求。文章通过面向对象的滑坡灾害特征提取所输出矢量结果可以方便地计算出滑坡的面积，对滑坡研究提供了极大的便利。%GF-2 satellite was successfully launched from Taiyuan, marking China remote sensing satellite into the submeter age and providing better remote sensing data for the investigation of geological disasters. GF-2 data preprocessing includes orthorectification, registration, fusion and clipping. Combining the spectral characterization, texture and DEM of landslide, this paper uses object-oriented classification method by seg-menting image, merging blocks, building classification rule and eliminating misclassification object to extract landslide information in Dongchuan District, Yunnan Province. The extraction result shows that, the method can meet the landslide recognition requirements with GF-2 data. The vector data extracted from landslide in-formation could be used to calculate landslide area and provide great convenience for landslide research.