大兴安岭火烧迹地遥感提取研究

Study on Remote Sensing Extractions of Burned Areas in Greater Khingan Mountains

林火是森林生态系统的重要影响因子,遥感技术的发展为林火监测和损失估算提供了强有力的技术手段。以内蒙古根河市1987年、2003年和2015年火烧迹地为研究对象,基于Landsat遥感影像,对比分析NDVI,EVI,GEMI,BAI,NBR,dNBR,NDSWIR等7种常见遥感指数对火烧迹地的提取能力,从而筛选出适合大兴安岭火烧迹地提取的最佳遥感指数。通过对遥感指数和分离指数的对比分析,得出结论:1)经历林火干扰的植被,其光谱特征也相应改变,基于近红外波段的遥感指数可将火烧迹地与正常植被区分。2)dNBR为提取大兴安岭火烧迹地的最佳遥感指数,其提取精度在90%以上;其次为NBR,BAI,NDSWIR;NDVI和EVI最差。3)基于dNBR提取的3个不同年份火烧迹地面积分别为3145.23,197726.67,48.06hm^(2)。

Fire is an important influencing factor of forest ecosystem.The development of remote sensing technology provides a powerful technical means for forest fire monitoring and loss estimation.Based on Landsat images,the extraction abilities of seven common remote sensing indexes(NDVI,EVI,GEMI,BAI,NBR,dNBR,NDSWIR)for the burned area in 1987,2003 and 2015 of Greater Khingan Mountains were compared and analyzed,so as to select the best remote sensing index suitable for the extraction of burning areas in Greater Khingan Mountains.Through the comparative analysis of the remote sensing indexes and the separation index,the following results were obtained:(1)after the fire,the spectral characteristics of vegetation changed correspondingly.The remote sensing index based on near-infrared band can distinguish the burned area from the normal vegetation.(2)The dNBR was the best remote sensing index for extracting the burned area in Greater Khingan Mountains,followed by the NBR,BAI and NDSWIR,the NDVI and EVI were the worst.(3)The areas of three burned areas based on dNBR were 3145.23,197726.67 and 48.06 hm^(2),respectively.