Fully Convolutional Network for Automatic Road Extraction from Satellite Imagery
This work addresses road extraction for traffic management, city planning, and road monitoring, but it is incremental as it builds on existing U-Net and ResNet architectures.
The paper tackles the problem of automatic and precise road extraction from high-resolution satellite imagery by proposing a fully convolutional neural network based on ResNet-34 and U-Net, achieving superior results in the DEEPGLOBE - CVPR 2018 road extraction sub-challenge with moderate memory usage enabling training on a single GTX 1080 or 1080ti.
Analysis of high-resolution satellite images has been an important research topic for traffic management, city planning, and road monitoring. One of the problems here is automatic and precise road extraction. From an original image, it is difficult and computationally expensive to extract roads due to presences of other road-like features with straight edges. In this paper, we propose an approach for automatic road extraction based on a fully convolutional neural network of U-net family. This network consists of ResNet-34 pre-trained on ImageNet and decoder adapted from vanilla U-Net. Based on validation results, leaderboard and our own experience this network shows superior results for the DEEPGLOBE - CVPR 2018 road extraction sub-challenge. Moreover, this network uses moderate memory that allows using just one GTX 1080 or 1080ti video cards to perform whole training and makes pretty fast predictions.