SDUNet: Road extraction via spatial enhanced and densely connected UNet

作者：

Highlights：

• We propose a network SDUNet, which combines the multi-level features of the road network and global prior information.

• A structure preserving model is proposed to enhance feature learning about the structure prior of the road surface.

• Experimental results show that our approach achieves state-of-the-art performance compared with previous methods.

摘要

•We propose a network SDUNet, which combines the multi-level features of the road network and global prior information.•A structure preserving model is proposed to enhance feature learning about the structure prior of the road surface.•Experimental results show that our approach achieves state-of-the-art performance compared with previous methods.

论文关键词：Road extraction,Image segmentation,Remote sensing imagery,Spatial topology

论文评审过程：Received 30 October 2020, Revised 8 November 2021, Accepted 22 January 2022, Available online 12 February 2022, Version of Record 16 February 2022.

论文官网地址：https://doi.org/10.1016/j.patcog.2022.108549