The land use classification based on Landsat 8 remote sensing imageA case study of Anqu demonstration community in Hongyuan County of Sichuan Province

According to the national standard of land use classification, based on the Landsat 8 OLI remote sensing data on July 23, 2013, combined with the ground observed data, digital elevation model, grassland resource type data and other auxiliary information, the visual interpretation signature of land use types were established in Anqu demonstration area in Hongyuan County of Sichuan Province. Regarding the results of the visual interpretation as the true values, we analyzed and compared the accuracy of unsupervised and supervised classifications and expert knowledge decision tree classification. The results show that, the supervised classification result has higher accuracy (85.02%) which is close to the actual situation comparing with other methods. The classification accuracy of the land use types of grassland, wetland, road and water, accounting for 99% of the entire study region, are 88.66%, 83.42%, 83.73% and 85.26%, respectively.