Скачать или смотреть 🚗 RoadLib: Instance-Level Road Marking Extraction, Mapping & Localization | Open-Source Demo

Welcome to the demo of RoadLib – a lightweight open-source library for visual road marking extraction, parameterization, mapping, and map-aided localization.

This project is an enhanced version of our RA-L 2022 work:
“Visual Mapping and Localization System Based on Compact Instance-Level Road Markings With Spatial Uncertainty”

🔹 What’s New in RoadLib?

✅ Batch pipeline ➡️ Incremental pipeline

✅ Ellipsoid ➡️ Bounding box parameterization

✅ High-precision poses ➡️ Local mapping + geo-registering

✅ Practical improvements for real-world datasets

🔹 Capabilities

Instance-level road marking segmentation (with a pretrained SegFormer model)

Incremental mapping & geo-registration

Map-aided localization using monocular images + odometry

Binary map outputs & Python-based visualization tools

🔹 Tested On:
Windows 10 & Ubuntu 20.04 (WSL2)

📂 Code & Documentation:
👉 GitHub Repository

📧 For training dataset details & collaboration, contact: [email protected]

✨ If you’re working on autonomous driving, SLAM, or road environment perception, this project can serve as a valuable reference.

👍 Don’t forget to Like, Share & Subscribe for more updates on mapping, localization, and robotics projects!