Large-scale Weakly Supervised Learning for Road Extraction from Satellite Imagery

TL;DR

This paper introduces a large-scale weakly supervised learning approach using OpenStreetMap data and satellite imagery to improve road extraction accuracy and generalization in diverse terrains and conditions.

Contribution

It leverages weak labels from OpenStreetMap and large-scale satellite data to pre-train models, surpassing current benchmarks and enhancing generalization across different regions.

Findings

Model trained on 100 times more data outperforms current top models.

Pre-training improves generalization to new areas.

Weakly supervised approach reduces need for pixel-level labels.

Abstract

Automatic road extraction from satellite imagery using deep learning is a viable alternative to traditional manual mapping. Therefore it has received considerable attention recently. However, most of the existing methods are supervised and require pixel-level labeling, which is tedious and error-prone. To make matters worse, the earth has a diverse range of terrain, vegetation, and man-made objects. It is well known that models trained in one area generalize poorly to other areas. Various shooting conditions such as light and angel, as well as different image processing techniques further complicate the issue. It is impractical to develop training data to cover all image styles. This paper proposes to leverage OpenStreetMap road data as weak labels and large scale satellite imagery to pre-train semantic segmentation models. Our extensive experimental results show that the prediction accuracy increases with the amount of the weakly labeled data, as well as the road density in the areas chosen for training. Using as much as 100 times more data than the widely used DeepGlobe road dataset, our model with the D-LinkNet architecture and the ResNet-50 backbone exceeds the top performer of the current DeepGlobe leaderboard. Furthermore, due to large-scale pre-training, our model generalizes much better than those trained with only the curated datasets, implying great application potential.