CCF DATA

The dataset was provided by the CCF Big Data Competition and consists of high-definition remote sensing images of a city in southern China in 2015. This dataset is relatively small and contains five large-sized RGB remote sensing images with annotations, ranging in size from 3000×3000 to 6000×6000. The annotations in the dataset classify four types of objects, including vegetation (marker 1), buildings (marker 2), water bodies (marker 3), roads (marker 4), and others (marker 0). we randomly crop small 256x256 images from these coordinates and perform data augmentation.