QinbaTownMorphologyDataset

This dataset comprises 16,128 manually annotated training images, each with a resolution of 512x512 pixels in PNG format. The images are sourced from high-resolution remote sensing imagery from Google Maps, specifically targeting small towns in the Qinba mountain area of southern Shaanxi. Utilizing advanced deep learning techniques and geometric algorithms, the dataset includes precisely defined building contours and town boundaries. It supports the identification of diverse town morphology types through comprehensive measurements and hierarchical cluster analysis. This dataset is designed for training U-Net models for image segmentation, aiming to enhance the understanding and analysis of urban morphology in complex topographical regions.