Building extraction method based on deep learning

The invention relates to the technical field of remote sensing image building extraction, in particular to a building extraction method based on deep learning, which comprises an encoder, a decoder and a semantic segmentation network model MFU-Net of a middle feature combination layer. The extraction of depth features is realized in the encoder stage; completing the recovery of the spatial resolution in the decoder stage; a middle-layer feature combination layer is added behind the decoder and is used for comprehensively considering different depth features and completing model output; when the model is trained, the final loss value of the model is calculated through summation of different losses, and compared with the prior art, extraction of deep features, recovery of spatial resolution and model training and output are achieved through a semantic segmentation network model of an encoder, a decoder and a middle-layer feature combination layer. According to the invention, buildings with different sizes can b