Building Type Classification Dataset from Satellite ImageryBuilding Type Classification Dataset from Satellite Imagery

This dataset contains 1,553 satellite images of buildings from across the United States, classified into 7 categories: Commercial (206), High-rise (294), Hospital (204), Industrial (217), Multi-family (220), Schools (206), and Single-family (206). Images are 512x512 pixels at approximately 0.15 m/pixel resolution, collected from Google Earth. The dataset was used for training a DenseNet201 model achieving 84.40% classification accuracy.