Abstract: Land water distribution is an indispensable component of global water resource security and management, climate research and ecological environment dynamic monitoring. Based on GF-1 data and Landsat8 OLI data, we used the mRMR feature selection algorithm and the object-oriented knowledge rule set to automatically extract the land water distribution of Hainan province in 2015, 2016 and 2017. Then assisted by high-resolution remote sensing images and Google Earth images, 500 sample points were selected to verify the classification accuracy. Results show a Kappa coefficient of 0.806, 0.869 and 0.913 for the three subsets, respectively, and both an error rate and an omission rate of lower than 0.1, demonstrating a high accuracy of the classification results. This dataset can be directly used to study the spatial and temporal distribution of land surface water. It also provides data basis for water environment research, such as water body studies and water resource safety assessment.
Keywords: land water; object-oriented knowledge rule set; Hainan province; medium and high resolution remote sensing data