基于环境卫星数据的森林叶面积指数遥感反演与验证――以大兴安岭加格达奇林区为例

以中国东北大兴安岭加格达奇林区为研究区，基于环境（HJ）卫星遥感数据提取森林植被指数，结合实测样点叶面积指数（leaf area index, LAI）数据构建研究区LAI遥感反演模型，获取研究区森林LAI。在此基础上，利用研究区LAI影像对LAI遥感产品GLOBCARBON LAI和MODIS LAI数据进行精度验证。研究结果表明:研究区LAI遥感反演模型中，基于比值植被指数（SR）的线性回归模型精度最高，模型R2为0.606（RMSE=0.251 6），相对误差19.89%；在研究区，GLOBCARBON LAI数据均值高于反演值，而MODIS LAI均值则相对较低，两者相对误差分别为12.2%和11.8%；通过对不同LAI值域的对比分析发现，研究区两种遥感LAI产品的最大误差均在LAI的低值区。
In this study, forest leaf area index (LAI) was mapped using LAI retrieved model based on remote sensing forest vegetation indexes from HJ satellite data and situ LAI measurements data in Jiagedaqi Distract, northeastern China. GLOBCARBON LAI and MODIS LAI products data were then validated using the retrieved LAI from HJ satellite. Results showed that the accuracy of LAI retrieved model based Simple Ratio (SR) was the highest with an R？？square of 0.606 and RMSE of 0.251 6. The relative error of retrieved LAI was 19.89% compared with measurement LAI data. In study area, the average value of GLOBCARBON LAI product was overestimated by 12.2% and the average value of MODIS LAI product was unde？？restimated by 11.8%. Furthermore, the errors in low values under 1.5 of the two LAI products were the largest according to comparison analysis of different value ranges of LAI