The Yangtse River_CIPs_10m_2021.rar

By combining random forest and active learning methods, a multi-dimensional discriminative feature space was constructed using a large number of collected and annotated samples from chemical industrial parks and non-chemical industrial park areas, along with publicly available Sentinel-2 remote sensing imagery data. This feature space integrates spectral and texture features, along with the general physical characteristics of chemical industrial parks identified through POI recognition. Utilizing partitioned training on the Google Earth Engine (GEE) platform, precise predictions of chemical industrial parks along the Yangtze River were achieved. Through technical validation, the model achieved an accuracy rate of 80%. This study provides, for the first time, distribution map data of chemical industrial parks within a 5km range along the Yangtze River with a resolution of 10m. It offers a scientific basis for the fine-grained governance of chemical industries along the Yangtze River and provides a model methodology guide for the precise identification and sustainable development of the chemical industry within world-class river basins.