基于环境减灾卫星遥感数据的呼伦贝尔草地地上生物量反演研究

推动国产遥感卫星在资源环境领域中的应用对于促进我国航天事业发展、减少科研成本具有重要意义。我国近期发射的环境减灾卫星具有时间分辨率高、可获得高光谱影像的特点,在陆地资源遥感监测领域将有广阔发展空间。研究于2009年夏季获得三景呼伦贝尔草原区遥感影像和对应地面实测草地生物量信息,基于这些数据探讨了利用环境减灾卫星多光谱影像和植被指数反演草地生物量的可行性。结果表明基于影像提取的NDVI、OSAVI、MSAVI、SAVI、EVI、MTVI2、WDRVI和GNDVI等光谱指数均与草地生物量有较好的定量关系。其中,MTVI2结果最好,预测决定系数达0.61,交叉检验决定系数为0.58,均方根误差仅为58.6 g.m-2,基于MTVI2和环境减灾卫星多光谱影像可准确生成草地生物量空间分布图。

Using Data of HJ-1A/B Satellite for Hulunbeier Grassland Aboveground Biomass Estimation

Promoting the application of domestic remote sensing satellite in natural resources and environment monitoring is very important to accelerate the development of national space research and reduce scientific activities cost. The recent launched satellite HJ-1A/B has the characteristic of high time resolution and can acquire hyperspectral image. These make it has the prosperous future in terrestrial resources detection. In the summer of 2009, the study obtained three images and corresponding grass dry biomass in Hulunbeier area. They were used to study the feasibility of using HJ-1A/B multispectral image and spectral indices for grassland biomass prediction. The result showed image based on calculation of spectral indices, including NDVI, OSAVI, MSAVI, SAVI, EVI, MTVI2, WDRVI, GNDVI, have good relationship with grass biomass. MTVI2 gained the best result with R² value of 0.61, meanwhile the cross validation result of it was also permissible with R² value of 0.58 and RMSE value of 58.6 g·m^-2. The map of grass biomass in the research area can be produced using MTVI2 and HJ multispectral imagery.