Road Extraction by Deep Residual U-Net

TLDR

Road extraction from aerial images has been a hot research topic in remote sensing image analysis. The study proposes a semantic segmentation neural network that merges residual learning with U‑Net to extract road areas from aerial images. The network, built with residual units and U‑Net–style skip connections, reduces training difficulty and parameter count while improving performance, and was evaluated on a public road dataset against U‑Net and other leading deep‑learning methods. The proposed model outperforms U‑Net and other state‑of‑the‑art road extraction methods, confirming its superiority.

Abstract

Road extraction from aerial images has been a hot research topic in the field of remote sensing image analysis. In this letter, a semantic segmentation neural network which combines the strengths of residual learning and U-Net is proposed for road area extraction. The network is built with residual units and has similar architecture to that of U-Net. The benefits of this model is two-fold: first, residual units ease training of deep networks. Second, the rich skip connections within the network could facilitate information propagation, allowing us to design networks with fewer parameters however better performance. We test our network on a public road dataset and compare it with U-Net and other two state of the art deep learning based road extraction methods. The proposed approach outperforms all the comparing methods, which demonstrates its superiority over recently developed state of the arts.