LoveDA

The LoveDA dataset contains high-spatial-resolution images from three different cities, focusing on improving the generalization capability of model from different urban and rural scenes.

BibTex: