Environment-Driven Online LiDAR-Camera Extrinsic Calibration

Abstract: LiDAR-camera extrinsic calibration (LCEC) is crucial for multi-modal data fusion in mechatronics. Existing methods, whether target-based or target-free, typically rely on customized calibration targets or fixed scene types, limiting their practicality in real-world applications. To address these challenges, we introduce EdO-LCEC, the first environment-driven online calibration approach. Unlike traditional target-free methods, EdO-LCEC observes the feature density of the application environment through a generalizable scene discriminator. Based on this feature density, EdO-LCEC extracts LiDAR intensity and depth features from varying perspectives to achieve higher calibration accuracy. To overcome the challenges of cross-modal feature matching between LiDAR and camera, we propose dual-path correspondence matching (DPCM), which leverages both structural and textural consistency for reliable 3D-2D correspondences. Additionally, our approach models the calibration process as a joint optimization problem utilizing global constraints from multiple views and scenes to enhance accuracy. Extensive experiments on real-world datasets demonstrate that EdO-LCEC outperforms state-of-the-art methods, particularly in sparse or partially overlapping sensor views.