基于Sentinel-2影像的火烧迹地提取及火烈程度分析OA

In recent years,under the combined influence of global climate change and human activities,forest fires have posed a severe threat to ecological security and sustainable development.Taking the 2023 forest fire in Jiuxi Town,Yuxi City,Yunnan Province as an example,using pre-fire and post-fire Sentinel-2 Level-1C(Sentinel-2 L1C)images,this study extracted the burned area information of Jiuxi Town utilizing the normalized difference vegetation index(INDV),a modified index based on the red edge band,the normalized burn ratio(RNB)based on the short-wave infrared band,and the spectral angle mapping(MSA)method.Additionally,the spatial distribution characteristics of fire severity were analyzed.The results are as follows.① The comparative analysis of INDV,RNB,and MSA indicates that the RNB difference method driven by short-wave infrared achieves the optimal extraction effect,with strong boundary continuity,and can accurately distinguish areas with different burn intensities.② The INDV1 index modified by the red edge band exhibits high sensitivity to vegetation dam-age among vegetation index methods,verifying its spectral advantages.③ The MSA method is greatly interfered by atmo-spheric and surface differences and is more suitable for extracting small-scale burned areas.④ The classification analysis of fire severity reveals that moderately damaged areas are dominant(accounting for 42％),and the fire spread from northeast to southwest,which is highly consistent with the monsoon characteristics of the study area.This study confirms the efficiency and practicality of Sentinel-2 data in fire assessment,providing scientific support for post-fire restoration and risk manage-ment in the ecologically fragile mountainous areas of Southwest China.