GLD-Road:A global-local decoding road network extraction model for remote sensing images
This work addresses efficient and accurate road extraction for applications like mapping and autonomous driving, representing an incremental improvement over existing methods.
The authors tackled road network extraction from remote sensing images by proposing GLD-Road, a two-stage model that combines global efficiency and local precision, resulting in improved APLS by 1.9% on City-Scale and 0.67% on SpaceNet3 datasets, while reducing retrieval time by 40% compared to Sat2Graph and 92% compared to RNGDet++.
Road networks are crucial for mapping, autonomous driving, and disaster response. While manual annotation is costly, deep learning offers efficient extraction. Current methods include postprocessing (prone to errors), global parallel (fast but misses nodes), and local iterative (accurate but slow). We propose GLD-Road, a two-stage model combining global efficiency and local precision. First, it detects road nodes and connects them via a Connect Module. Then, it iteratively refines broken roads using local searches, drastically reducing computation. Experiments show GLD-Road outperforms state-of-the-art methods, improving APLS by 1.9% (City-Scale) and 0.67% (SpaceNet3). It also reduces retrieval time by 40% vs. Sat2Graph (global) and 92% vs. RNGDet++ (local). The experimental results are available at https://github.com/ucas-dlg/GLD-Road.