Abstract: This study draws sources from 86-scene Landsat 8 OLI satellite images with cloud volume less than 10%, high-temporal resolution GF-1 WFV wide-viewing angle multispectral camera, UAV (unmanned aerial vehicle) images, and Google Earth images. According to the the topographical and land cover features of Henan Province, four sample regions with a full range of land cover types were selected for conducting a comprehensive field survey, through which we obtained 1518 samples. Then ArcMap 10.3 was used to visually interpret the sample area (to the secondary type) based on the characteristics of sample points and land cover features. Experts were asked to evaluate and correct the interpretation results. On the basis of the interpretation of sample area, another field survey was conducted in the remaining land of Henan Province, and a total of 4,883 field samples were obtained at last. Finally, 1,895 ground verification points were selected through the global random regular hexagon grid (12 levels) method to evaluate the interpretation accuracy. The results showed that the overall interpretation accuracy of the land cover in Henan Province was 91.19%, and the Kappa coefficient was 0.886.
Keywords: Henan Province; land cover; Landsat 8; visual interpretation; 2015