基于GEE的林火识别与火后植被恢复评估研究
收藏资源简介：
基于GEE获取根河市2001~2010年的年际TM合成影像，选择以影像的近红外波段与 dNBR指数作为过火区敏感特征，利用OTSU 算法计算阈值后构建决策树分类模型提取根河市年际火烧迹地，验证精度后建立了市级林火历史属性数据库，分析森林火灾的时空分布情况；根据 dNBR 指数阈值范围进行林火烈度分级，结合DEM 数据分析林火烈度的空间分布；基于 GEE 调用MOD09A1数据集获取过火区及邻近植被区斑块的NDVI、NBR 指数时间序列数据，通过对比分析确定火烧迹地起火时间；基于GEE调用Landsat系列数据合成的EVI指数产品提取不同火烧迹地各级林火烈度区域像元的EVI指数时间序列数据，以起火前三年生长期的EVI均值为对照计算dEVI指数，从时间尺度上分析不同林火烈度对火后植被恢复的影 响，评估火后植被恢复年限。
Based on Google Earth Engine (GEE), annual TM composite images of Genhe City from 2001 to 2010 were collected. The near-infrared band and delta Normalized Burn Ratio (dNBR) index of the images were selected as sensitive features for burned areas. After calculating the optimal threshold using Otsu's algorithm, a decision tree classification model was built to extract annual burned areas in Genhe City. Following accuracy verification, a municipal forest fire historical attribute database was established, and the spatiotemporal distribution of forest fires was analyzed. Forest fire severity was classified based on the threshold range of the dNBR index, and the spatial distribution of fire severity was investigated combined with Digital Elevation Model (DEM) data. Using GEE to call the MOD09A1 dataset, time series data of Normalized Difference Vegetation Index (NDVI) and Normalized Burn Ratio (NBR) for patches in burned areas and adjacent vegetation regions were acquired, and the ignition time of burned areas was determined through comparative analysis. Using GEE to call the Enhanced Vegetation Index (EVI) products synthesized from Landsat series data, time series data of EVI index for pixels in regions with different fire severity levels of various burned areas were extracted. The dEVI index was calculated by taking the mean EVI value of the three growing seasons before ignition as the control group. The impact of different fire severity levels on post-fire vegetation restoration was analyzed from the temporal perspective, and the post-fire vegetation recovery period was evaluated.
