基于HJ-1高光谱影像的黄河口芦苇和碱蓬生物量估测模型研究

湿地植被的生物量是湿地生态评价、保护和利用的重要基础数据,遥感技术已经成为湿地生物量高效、准确监测的重要手段。基于2013年9月的HJ-1 高光谱遥感影像,应用准同步现场踏勘数据,通过单变量线性回归和多变量线性回归的方法,针对7种常用的窄波段植被指数和2种红边指数对黄河口芦苇和碱蓬生物量(地上干重)的估测能力进行了评价。结果表明:(1)单光谱指数变量情况下,对于芦苇,选择近红外827 nm波段和红635 nm波段简单植被指数(SRI)和线性插值红边指数(REP_ linear interpolation)取得了最佳的单变量回归结果,决定系数分别达到0.42和0.58;对于碱蓬,选择近红外807 nm波段和红692 nm波段的归一化差值植被指数(NDVI)、SRI和优化的土壤校正植被指数(OSAVI)取得了较好的回归结果,决定系数分别达到0.60,0.59和0.47;(2)多光谱指数变量情况下,以在单变量回归分析中取得较好结果的SRI和REP_ linear interpolation指数为变量,芦苇得到了与其生物量之间决定系数为0.71的高相关性;同时,以NDVI、SRI和OSAVI为变量,与碱蓬生物量的决定系数达到了0.66。

Reeds and suaeda biomass estimation model based on HJ-1 hyperspectal image in the Yellow River Estuary

Wetland vegetation biomass is the basic information of wetland ecological assessment, protection and utilization. Remote sensing has become one of the most efficient technologies of wetland biomass monitoring. Utilizing the HJ-1 hyperspectral remote sensing image that acquired in September 2013 and the coinstantaneous field survey data, the biomass estimation capabilities of 7 kinds of narrow-band vegetation indices and 2 kinds of red edge position indices of reeds and suaeda in the Yellow River Estuary have been studied. The results reveal that (1) In single variable estimation model case, for reeds, the SRI index with 635 nm and 827 nm bands and REP_ linear interpolation index get the best R² measures, and for suaeda, the NDVI and SRI indices with 692 nm and 807 nm bands and OSAVI index get the best biomass estimation results. (2) In multiple variable case, for reeds and suaeda, the R² measure get 0.71 and 0.66 respectively.