Building Footprint Extraction with Graph Convolutional Network
This addresses the challenge of accurate boundary delineation in urban modeling for applications like 3-D reconstruction, though it appears incremental as it builds on existing deep learning approaches.
The paper tackled the problem of precise building footprint extraction from satellite images by proposing an end-to-end framework using a graph convolutional network (GCN), which outperformed state-of-the-art methods.
Building footprint information is an essential ingredient for 3-D reconstruction of urban models. The automatic generation of building footprints from satellite images presents a considerable challenge due to the complexity of building shapes. Recent developments in deep convolutional neural networks (DCNNs) have enabled accurate pixel-level labeling tasks. One central issue remains, which is the precise delineation of boundaries. Deep architectures generally fail to produce fine-grained segmentation with accurate boundaries due to progressive downsampling. In this work, we have proposed a end-to-end framework to overcome this issue, which uses the graph convolutional network (GCN) for building footprint extraction task. Our proposed framework outperforms state-of-the-art methods.