An Improved Method for Road Extraction from High-Resolution Remote-Sensing Images that Enhances Boundary Information

Abstract

At present, deep-learning methods have been widely used in road extraction from remote-sensing images and have effectively improved the accuracy of road extraction. However, these methods are still affected by the loss of spatial features and the lack of global context information. To solve these problems, we propose a new network for road extraction, the coord-dense-global (CDG) model, built on three parts: a coordconv module by putting coordinate information into feature maps aimed at reducing the loss of spatial information and strengthening road boundaries, an improved dense convolutional network (DenseNet) that could make full use of multiple features through own dense blocks, and a global attention module designed to highlight high-level information and improve category classification by using pooling operation to introduce global information. When tested on a complex road dataset from Massachusetts, USA, CDG achieved clearly superior performance to contemporary networks such as DeepLabV3+, U-net, and D-LinkNet. For example, its mean IoU (intersection of the prediction and ground truth regions over their union) and mean F1 score (evaluation metric for the harmonic mean of the precision and recall metrics) were 61.90% and 76.10%, respectively, which were 1.19% and 0.95% higher than the results of D-LinkNet (the winner of a road-extraction contest). In addition, CDG was also superior to the other three models in solving the problem of tree occlusion. Finally, in universality research with the Gaofen-2 satellite dataset, the CDG model also performed well at extracting the road network in the test maps of Hefei and Tianjin, China.

Automatic Road Extraction from High-Resolution Remote Sensing Images Using a Method Based on Densely Connected Spatial Feature-Enhanced Pyramid

“…This paper proposes an improved semantic segmentation model that implements automatic road extraction of highresolution remote sensing images. To verify the performance of this method, we compared the DGRN with four representative deep learning network models of U-net [38], DeepLabV3+ [42], D-LinkNet [43] and CDG [39] on two data sets under the same conditions. The U-net network model is mainly used for medical image segmentation.…”

“…Zhang et al [37] combined ResNet and Unet [38] to extract road networks, which are designed with few parameters but achieve good performance. Recently, Wang et al [39] proposed coord-dense-global model(CDG), which uses DenseNet, the coordconv module and the global attention module to construct the road map from remote sensing imagery. Although these FCN-based methods perform well in road extraction, they perform poorly when nonroad objects occlude the road.…”

Generative Learning for Postprocessing Semantic Segmentation Predictions: A Lightweight Conditional Generative Adversarial Network Based on Pix2pix to Improve the Extraction of Road Surface Areas

“…In [37], the authors extract the linear characteristics of selected road segments by constructing a geometric knowledge base of rural roads to improve the initial results. In [38], a network for road extraction with a final module to highlight high level information and improve the classification is proposed.…”

Cascaded Residual Attention Enhanced Road Extraction from Remote Sensing Images

“…The performance of road extraction was improved using dense blocks in [29,35]. The Coord-Dense-Global (CDG) model [36] introduced an attention mechanism to enhance the edge information and global context of roads based on DenseNet. For road boundary refinement, Conditional Random Fields (CRF) were introduced as post-processing strategies to optimize the extracted result [37,38].…”