GF-1 WFV与Landsat 8 OLI的NDVI交互比较与定量关系研究

Inter-Sensor Comparison and Quantitative Relationships Between GF-1 WFV and Landsat 8 OLI NDVI Data

为弥补、修正构建Landsat 8归一化植被指数(normalized differential vegetation index,NDVI)长时间序列中因云层覆盖、影像质量与重访周期导致的数据缺失问题,交互比较高分一号(GF-1)与Landsat 8两类传感器的植被探测能力,并对其植被指数的定量关系进行研究,探讨以GF-1局部镶嵌或全局替换Landsat 8缺失数据构建时序NDVI的可行性。结果显示,Landsat 8与GF-1具有NDVI差异性,具体表现为Landsat 8探测地物信号更强,而GF-1获取地物信息量更多;相同空间分辨率(30 m)下两传感器NDVI数据呈现高度线性相关,经方程转换后,GF-1的NDVI拟合数据与Landsat 8的NDVI关系增强;同时,两传感器的NDVI数据存在区域差异性,且拟合数据可以有效减少差异程度。实验表明,此方法可实现Landsat 8高分辨率时序NDVI的构建,具有一定应用价值。

In order to compensate or correct the missing data caused by cloud cover, image quality or revisiting period in the process of building long time normalized differential vegetation index(NDVI) series of Landsat 8, the vegetation detectability and quantitative relationship of NDVI between GF-1 and Landsat 8 were researched, and the feasibility to construct time series NDVI by replacing the missing data from Landsat 8 with GF-1 through local or global mosaicking was further investigated. The result reveals differences between two sensors in NDVI, specifically Landsat 8 detects ground signals more powerful and GF-1 gets more information about objects;The NDVI data of Landsat 8 and GF-1 under 30 m resolution follow the linear fitting, and the regression equation enhanced the relationship between two sensors after conversation;Meanwhile, the NDVI values between two sensors exist regional differences, and the fitting data can effectively reduce the difference.The experimental results show that this method can realize the construction of high resolution NDVI time series of Landsat 8 with certain application value.