摘要
针对森林火灾火烧迹地提取精度不够高的问题,基于Sentinel-2等新型卫星提供的对植被敏感的红边波段,分别构建了基于主成分分析法和新型植被指数的火烧迹地提取算法。第一种是基于第三主成分分析法(PC3),选择可见光及近红外波段进行主成分分析,根据变换后的第三主成分提取火烧迹地,将3个红边波段分别替换可见光红波段,根据得到的PC3red-dege1、PC3red-dege2、PC3red-dege3提取火烧迹地。第二种是基于新型植被指数法,根据NDVI及将3个红边波段分别替换可见光红波段计算获取的NDVIre1、NDVIre2、NDVIre3,分别提取火烧迹地。以2017年毕拉河特大火灾火烧迹地为研究区。实验结果表明,基于PC3的火烧迹地取效果整体优于基于植被指数的提取效果;红边波段参与的火烧迹地的提取精度高于红波段参与的提取精度,其中PC3red-dege1精度最高,PC3red-dege2提取精度次之。研究证明利用红边波段采用PC3的方法能满足火烧迹地的高精度快速提取。
Aiming at the problem of low accuracy in extracting forest fire burns,based on vegetation-sensitive Red-edge bands provided by new satellites such as Sentinel-2,this paper constructed two fire burned ground extraction algorithms based on principal component analysis and new vegetation index.The first is based on the third principal component analysis(PC3).It selects the visible light and NIR wavebands for principal component analysis,and extracts the fire burned ground based on the transformed third principal component,while replacing the R band of visible light with the three Red-edge wavebands,respectively,and based on the obtained PC3 red-dege1 ,PC3 red-dege2 ,and PC3 red-dege3 ,the fire burned ground is extracted.The second is based on the new vegetation index.Based on NDVI and the NDVI re1 ,NDVI re2 and NDVI re3 obtained by replacing the visible R band with the three Red-edge bands,respectively,it extracts the fire burned ground.This paper takes the fire area of the Biya river in 2017 as the research area.The experimental results show that:the overall extraction effect of PC3-based is better than that based on vegetation index;the extraction precision of the fire burned ground involved in the Red-edge band is higher than the extraction precision of the R band participated,of which the PC3 red-dege1 precision is the highest and the PC3 red-dege2 is the second.The research proves that the method of PC3 using Red-edge band can meet the high accuracy and fast extraction of the fire burned ground.
作者
李莹
于海洋
王燕
李朝亮
江一帆
LI Ying;YU Haiyang;WANG Yan;LI Chaoliang;JIANG Yifan(Key Laboratory of Mine Spatial Information Technologies of NASG,Henan Polytechnic University,Jiaozuo,Henan 454000,China)
出处
《遥感信息》
CSCD
北大核心
2019年第5期63-68,共6页
Remote Sensing Information
基金
国家自然科学基金（U1304402）
国家重点研发计划（2016YFC08033103）
河南理工大学青年骨干教师项目（672105/216）
