Effect of spatial and spectral resolution of images on interpreting intertidal estuarine sediment grain size distributions

摘要

In this paper, multi- and hyper- spectral images were employed to evaluate the capacity of retrieving grain size of tidal flat sediment from space-borne remote sensing data by analyzing in situ spectral signatures and grain size parameters. The results show that the mean absolute error (MAE) of medium grain size retrieved from HJ-1A hyperspectral in wet tidal flat with a moisture content of 18-43% is 0.0036. In dry tidal flat, the MAE of medium grain size retrieved from Landsat ETM and HJ-1A images are 0.031 and 0.026 respectively. Landsat MSS can stinguish sediment types of FS, TS and ST. Contrastive analysis of three kind of images, Landsat MSS is effective in historic period large scale dry tidal flat, and HJ-1A image has high precision in recent year's wet tidal flat sediment types.