激光生物学报, 2014, 23 (6): 604, 网络出版: 2015-07-07
基于高分一号 WFV卫星影像的黄河口湿地草本植被生物量估算模型研究
Study on the Above Ground Vegetation Biomass Estimation Model Based on GF-1 WFV Satellite Image in the Yellow River Estuary Wetland
摘要
基于高分一号WFV卫星数据, 通过计算NDVI、SRI、SAVI、MSAVI、GBNDVI和DVI 6种植被指数, 结合黄河口湿地现场草本植被地上干生物量样本, 进行一元线性、指数、对数和幂4种模型的回归分析, 从中选出生物量最佳估算模型。结果表明: (1)指数模型是NDVI和GBNDVI的最佳生物量回归模型, 幂模型是其它植被指数的最佳回归模型(P<0.001); (2)6种植被指数干生物量最佳回归模型的决定系数(R2)都大于0.7, 最大为0.77, 由大到小排序为MSAVI>SAVI>NDVI>DVI>SRI>GBNDVI; (3)各最佳回归模型的平均相对误差(MRE)都小于54%, 最小为23.9%, 由小到大排序为NDVI
Abstract
Six kinds of vegetation indices, including NDVI, DVI, SRI, SAVI, MSAVI and GBNDVI were calculated from GF-1 WFV satellite image of the Yellow River estuary, These indices were regressed with the wetland aboveground herbaceous vegetation dry biomass data derevied from in situ sampling. The linear, exponential, logarithmic and power functions were adopted for the regression analysis respectively. The best biomass estimation models wer determined: (1)The best biomass estimation models based on NDVI and GBNDVI were exponential regression model, and the models based on other vegetation indices were power regression model(P<0.001); (2) The determination coefficients for all of the best regression models’ were higher than 0.7, and the highest one was 0.77. The sequence of them was MSAVI>SAVI>NDVI>DVI>SRI>GBNDVI. (3) The MRE of the best regression models based on these vegetation indices were less than 54%, and the minimum was 23.9%. The sequence of them was NDVI
王建步, 张杰, 马毅, 任广波. 基于高分一号 WFV卫星影像的黄河口湿地草本植被生物量估算模型研究[J]. 激光生物学报, 2014, 23(6): 604. WANG Jianbu, ZHANG Jie, MA Yi, REN Guangbo. Study on the Above Ground Vegetation Biomass Estimation Model Based on GF-1 WFV Satellite Image in the Yellow River Estuary Wetland[J]. Acta Laser Biology Sinica, 2014, 23(6): 604.
PDF全文
