Extracting Information on Wetlands in Northeast China Based on Time Series FY3/MERSI Data

Abstract

Using Fengyun Medium Resolution Spectral Imager (FY3/MERSI) data from 2011 to 2018, this paper analyzed the characteristics of original spectra and the normalized difference vegetation index (NDVI) of lakes, reservoirs, rivers, paddy fields and swamps in Northeast China, A large-scale wetland extraction model was built using the decision tree classification method, and accuracy was verified in typical sites using China High-resolution Earth Observation System (CHEOS-1) data. The results showed that there were differences in the spectral curves and NDVI of different wetland types, especially the swamps, paddy fields and other landforms that showed a growth cycle with more significant differences in spectral curves and index curves than those of lakes, reservoirs and rivers. The total area of wetlands in Northeast China was 79,123.4 km2, including 5262.3 k m2 of lake and reservoir wetlands, 6514.7 km2 of river wetlands, 15284.7 k m2 of swamp wetlands and 52061.7 km2 of paddy fields. The accuracy verification showed that the area accuracy of wetlands was over 91.8%, the overall classification accuracy of wetlands was over 84.8% and the Kappa coefficient was over 0.6446.