A Residual Attention and Local Context-Aware Network for Road Extraction from High-Resolution Remote Sensing Imagery

Ziwei Liu,Mingchang Wang,Fengyan Wang,Xue Ji

doi:10.3390/rs13244958

Abstract

Extracting road information from high-resolution remote sensing images (HRI) can provide crucial geographic information for many applications. With the improvement of remote sensing image resolution, the image data contain more abundant feature information. However, this phenomenon also enhances the spatial heterogeneity between different types of roads, making it difficult to accurately discern the road and non-road regions using only spectral characteristics. To remedy the above issues, a novel residual attention and local context-aware network (RALC-Net) is proposed for extracting a complete and continuous road network from HRI. RALC-Net utilizes a dual-encoder structure to improve the feature extraction capability of the network, whose two different branches take different feature information as input data. Specifically, we construct the residual attention module using the residual connection that can integrate spatial context information and the attention mechanism, highlighting local semantics to extract local feature information of roads. The residual attention module combines the characteristics of both the residual connection and the attention mechanism to retain complete road edge information, highlight essential semantics, and enhance the generalization capability of the network model. In addition, the multi-scale dilated convolution module is used to extract multi-scale spatial receptive fields to improve the model’s performance further. We perform experiments to verify the performance of each component of RALC-Net through the ablation study. By combining low-level features with high-level semantics, we extract road information and make comparisons with other state-of-the-art models. The experimental results show that the proposed RALC-Net has excellent feature representation ability and robust generalizability, and can extract complete road information from a complex environment.

Highlights

The rapid development of remote sensing technology has led to a sharp increase in high-resolution remote sensing image data and provides an important data source for road extraction
RALC-Net was compared with four state-of-the-art road extraction methods based on deep learning models, including SegNet [64], U-Net [56], DeeplabV3+ [57], and D-LinkNet [65], which compare the extraction results of two input ways using only image data and using multi-feature information
The experimental results verified the challenging performance of the RA module

Summary

Introduction

The rapid development of remote sensing technology has led to a sharp increase in high-resolution remote sensing image data and provides an important data source for road extraction. Road extraction methods are divided into traditional methods and deep learning-based methods [9,10]: (1) Traditional methods In this method, road types are defined manually based on one particular feature information, and the corresponding feature extraction model is constructed to recognize and extract the road [11,12]; (2) Deep learning-based methods. Road types are defined manually based on one particular feature information, and the corresponding feature extraction model is constructed to recognize and extract the road [11,12]; (2) Deep learning-based methods This method takes advantage of automatic feature extraction ability, strong generalization ability, high efficiency of fitting ability, and stability robustness of the deep learning model to mine deep characteristic information from HRI based on the prior knowledge to complete the automatic extraction of road information [13,14,15]

Results

Discussion

Conclusion