2020年度一带一路滑坡、泥石流多发区灾害监测产品

利用随机森林模型和Google Earth Engine云平台合成的年度Landsat 8数据为基础，构建了滑坡提取模型，提取出了青藏高原南麓范围内2020年滑坡信息二值图，横断山脉2020年滑坡信息二值图。将每个像素在相邻时相年度合成的Landsat 8数据的各个波段值和计算的NDVI, NDMI, NDSI, NBR和斜率作为数据特征，训练基于像素的滑坡提取模型。模型结构采用随机森林，训练数据以2015年Nepal Gorkha大地震发布的滑坡解译结果为基准。最后以二值图的形式输出滑坡提取结果（滑坡的像素值为1）。

Based on the annual composite Landsat 8 data generated via the Google Earth Engine (GEE) cloud platform, a landslide extraction model was built using the Random Forest algorithm. Two binary maps of 2020 landslide information were extracted, covering the southern foothills of the Qinghai-Tibet Plateau and the Hengduan Mountains, respectively. Various band values of each pixel from the annual composite Landsat 8 data across adjacent temporal phases, along with the calculated NDVI, NDMI, NDSI, NBR and terrain slope, were utilized as data features to train the pixel-based landslide extraction model. The model adopted the Random Forest architecture, and the training dataset was referenced against the official landslide interpretation results released after the 2015 Nepal Gorkha Earthquake. Finally, the landslide extraction results were output in the form of a binary map, where the pixel value of landslide-covered areas was set to 1.

搜集汇总

背景与挑战

背景概述

该数据集为2020年度一带一路滑坡、泥石流多发区灾害监测产品，基于Landsat 8数据和随机森林模型提取了青藏高原南麓和横断山脉的滑坡信息，输出为二值图格式，空间分辨率30m。