论文部分内容阅读
受稀疏植被与明亮土壤背景影响,干旱地区植被覆盖精确遥感估测难度较大。以Hyperion影像为数据源,选取甘肃省民勤绿洲-荒漠过渡带为研究区,系统比较了利用不同类型高光谱及多光谱植被指数估测干旱地区稀疏植被覆盖度的能力,以期确定干旱地区稀疏植被覆盖度估测的最佳植被指数。不同植被指数估测稀疏植被覆盖度的能力利用线性回归R2及留一交叉验证的均方根误差进行比较,结果表明:高光谱植被指数估测稀疏植被覆盖度的能力显著优于相应的多光谱植被指数,抗大气植被指数(ARVI)及抗土壤和大气植被指数(SARVI)表现明显优于归一化植被指数(NDVI)与土壤调节植被指数(SAVI),其中以基于833.3nm/640.5nm波段组合的ARVI表现最佳,R2可达0.7294,均方根误差(RMSE)仅为5.5488。
Affected by sparse vegetation and bright soil background, it is very difficult to accurately estimate the vegetation coverage in arid areas. Taking Hyperion as data source and Minqin Oasis - Desert Transitional Zone of Gansu Province as the research area, the ability of using different types of hyperspectral and multispectral vegetation indices to estimate sparse vegetation coverage in arid regions was systematically compared in order to determine the sparseness of arid regions Vegetation coverage estimation of the best vegetation index. The ability of different vegetation indices to estimate the sparse vegetation coverage was compared using linear regression R2 and root mean square error with a cross-validation. The results showed that hyperspectral vegetation index’s ability to estimate sparse vegetation coverage was significantly better than that of corresponding multi- Vegetation index, anti-atmospheric vegetation index (ARVI) and anti-soil and atmospheric vegetation index (SARVI) performance was significantly better than the normalized NDVI and SAVI, which based on 833.3nm / 640.5nm band The combined ARVI performed best with an R2 of 0.7294 and a root mean square error (RMSE) of only 5.5488.