Deep Learning Code for Building Extraction from Very high resolution (VHR) remote sensing images

The main architecture is the CMGFNet architecture that was used for multimodal image fusion, this model has been successful in our tasks such as building extraction from VHR.

The main dataset used to train the network was the Postdam and Vaihingen Dataset created by ISPRS Community. Below is an example of a training image overlayed with the Building and correspending digital surface model mask.