基于无人机大样方数据及国产卫星反演草地植被覆盖度方法研究

受开垦、采矿等人类活动影响，草原出现退化甚至沙化，监测其植被覆盖度对于揭示草地的分布状况与空间变化规律具有重要意义。本文以无人机大样方数据与国产高分一号（GF-1）数据作为数据源，结合野外同步数码相机获取的数据，应用支持向量机（Support vector machine，SVM）构建不同数据源之间的植被覆盖度反演模型（数码相片—无人机大样方数据植被覆盖度估算模型，无人机大样方数据—GF-1数据植被覆盖度估算模型），探讨国产GF-1卫星结合无人机大样方估算草原植被覆盖度的方法。结果表明，基于SVM模型的GF-1数据结合无人机大样方计算的土壤调节植被指数（Soil-adjusted vegetation index，SAVI）具有较高的精度（判定系数R²=0.97，相对分析误差RPD=4.86，均方根误差RMSE=3.23），因此基于无人机大样方数据结合GF-1数据可以准确、快速地反演草地覆盖度，利用这种方法可以估算整个草原的植被覆盖度。

Research on Grassland Fractional Vegetation Coverage Inversion Method Based on Drone Large Quadrat Data and Domestic Satellite

Hulun Buir Prairie has undergone some desertification and degradation due to the overgrazing,farming and mining. It is of great significance to monitor grassland coverage of Hulun Buir Prairie to uncover the distribution and spatial variety regulation of grassland. The Support Vector Machine (SVM) models is established based on the different vegetation indices from GF-1 multi-spectral data and ground data from unmanned aerial vehicle (UAV)to retrieve grassland coverage. The aim of this study is to evaluate the potential ability of Chinese GF-1 satellite imagery combining with UAV photogrammetry in grassland FVC retrieval at the surface mine area of north prairie in China. The results showed that soil-adjusted vegetation index (SAVI) from GF-1 can produce high accuracy estimation (R²=0.956,RPD=4.857,RMSE=3.232) based on SVM model. Therefore,the Chinese GF-1 data can provide grassland coverage with high accuracy based on SVM model. It was found that this method was spatially consistent,allowing accurate vegetation mapping over the entire grassland.