Remote Sensing Extraction of Agricultural Land in Shandong Province, China, from 2016 to 2020 Based on Google Earth Engine

Timely and effective access to agricultural land-change information is of great significance for the government when formulating policies. Due vast area Shandong Province, current research on land use in Province very limited. The classification accuracy methods also needs be improved. In this paper, with support Google Earth Engine (GEE) platform based Landsat 8 time series image data, a multiple machine learning algorithm was used obtain spatial variation distribution from 2016 2020. Firstly, high-quality cloud-free synthetic dataset 2020 obtained using GEE. Secondly, thematic index calculated phenological characteristics land, periods significant differences terms water, artificial surface, woodland bare were selected classification. Feature information, such as texture features, spectral features terrain constructed, random forest method select optimize features. Thirdly, forest, gradient boosting tree, decision tree ensemble algorithms classification, four classifiers compared. changes extracted causes analyzed. results show following: (1) multi-spatial more accurate than single obtaining characteristics; (2) classifier. overall five land-extraction by above 0.9; (3) annual decrease related increase land-surface urbanization rate.