FAST-Calib: LiDAR-Camera Extrinsic Calibration in One Second

2025-07-23

AI Summary

To address challenges in LiDAR–camera extrinsic calibration—including difficult cross-modal registration, point cloud edge blurring degrading accuracy, and limited generality across solid-state and mechanical LiDARs—this paper proposes a fully automatic calibration method leveraging a custom-designed 3D calibration target. Our approach introduces three key innovations: (1) a scan-pattern-agnostic edge extraction algorithm for robust 3D edge detection; (2) ellipse fitting to compensate for edge expansion caused by laser spot diffusion; and (3) a multi-scene joint nonlinear optimization framework enhancing robustness and generalizability. Evaluated on three distinct LiDAR types, the method achieves a point-to-point reprojection error ≤6.5 mm and completes calibration in ≤0.7 s per instance, outperforming state-of-the-art methods in both accuracy and stability. The source code and dataset are publicly available.

Technology Category

• Computer Vision: 3D Computer Vision
• Intelligent Robots: Multimodal Perception & Sensor Fusion
• Machine Learning: Calibration & Uncertainty Quantification

Application Category

• Search and Retrieval-Augmented AI: Retrieval-Augmented Generation (RAG) and multi-modal RAG
• Systems and Infrastructure for Web, Mobile and WoT: Experiences and lessons learnt from Web-based algorithms and system deployments
• Graph Algorithms and Modeling for the Web: Efficient manipulation of static and dynamic Web-related graphs

Abstract

This paper proposes FAST-Calib, a fast and user-friendly LiDAR-camera extrinsic calibration tool based on a custom-made 3D target. FAST-Calib supports both mechanical and solid-state LiDARs by leveraging an efficient and reliable edge extraction algorithm that is agnostic to LiDAR scan patterns. It also compensates for edge dilation artifacts caused by LiDAR spot spread through ellipse fitting, and supports joint optimization across multiple scenes. We validate FAST-Calib on three LiDAR models (Ouster, Avia, and Mid360), each paired with a wide-angle camera. Experimental results demonstrate superior accuracy and robustness compared to existing methods. With point-to-point registration errors consistently below 6.5mm and total processing time under 0.7s, FAST-Calib provides an efficient, accurate, and target-based automatic calibration pipeline. We have open-sourced our code and dataset on GitHub to benefit the robotics community.

Problem

• Fast LiDAR-camera extrinsic calibration using a 3D target
• Edge extraction algorithm for various LiDAR scan patterns
• Compensating LiDAR spot spread artifacts via ellipse fitting

Innovation

• Uses custom-made 3D target for calibration
• Edge extraction algorithm agnostic to LiDAR
• Compensates edge dilation via ellipse fitting