基于Sentinel-2影像的SAVI快速过火区面积提取

Extraction of forest burned area using SAVI based on Sentinel-2

森林大火严重影响生态环境及人民生命财产安全。快速准确的提取过火迹地范围、掌握火灾情况,快速做出应急响应非常重要。本文以2020年3月30日四川凉山州森林大火为例,利用灾前、灾后Sentinel-2数据,采用影像归一化燃烧比(NBR)和土壤调整植被指数(SAVI)对火烧区域进行检测,对差分归一化燃烧比(dNBR)和差分土壤调整植被指数(dSA VI)进行非监督分类,快速提取过火区面积。研究结果表明,利用dSAVI识别的过火区面积更接近官方数据,火烧迹地识别精度达到92.83%,且Sentinel-2数据利用SAVI指数获取实时过火面积的方法简单快捷、所提取的结果更准确,其可以为灾害评估和灾后重建提供参考,为森林火灾应急响应提供技术支撑。

Forest fire seriously affects the ecological environment and the safety of people’s life and property.Therefore,it is very important to extract the area of fire mark quickly and accurately,master the serious situation of fire and make emergency response quickly.In this paper,the forest fire in Liangshan Prefecture,Sichuan province on March 30,2020 was taken as an example.The sentinel-2 data before and after the fire were used as the research object,and normalized burning ratio(NBR) and soil-adjusted vegetation index(SAVI) were used to detect the burned area.Differential normalized combustion ratio(dNBR) and differential soil-adjusted vegetation index(dSAVI) were classified unsupervised,and the area of overburning area was extracted quickly.The results show that the area of the fire zone identified by dSAVI is closer to the official data,and the extraction accuracy reaches 92.83%.The method to obtain real-time burned area by using SAVI index for Sentinel-2 data is simple and fast,and the extracted results are accurate enough.The research results can provide reference for disaster assessment and post-disaster reconstruction,and provide technical support for forest fire emergency response.