SOIC: Semantic Online Initialization and Calibration for LiDAR and Camera
This addresses the calibration challenge for autonomous vehicles and robotics by providing an incremental improvement over existing online methods that require initial guesses.
The paper tackles the problem of online extrinsic calibration for LiDAR and camera sensors by proposing SOIC, a semantic-based approach that eliminates the need for prior rough initial values, converting initialization to a Perspective-n-Point problem and refining parameters with a semantic constraint cost function, with experimental validation on the KITTI dataset showing feasibility and accuracy.
This paper presents a novel semantic-based online extrinsic calibration approach, SOIC (so, I see), for Light Detection and Ranging (LiDAR) and camera sensors. Previous online calibration methods usually need prior knowledge of rough initial values for optimization. The proposed approach removes this limitation by converting the initialization problem to a Perspective-n-Point (PnP) problem with the introduction of semantic centroids (SCs). The closed-form solution of this PnP problem has been well researched and can be found with existing PnP methods. Since the semantic centroid of the point cloud usually does not accurately match with that of the corresponding image, the accuracy of parameters are not improved even after a nonlinear refinement process. Thus, a cost function based on the constraint of the correspondence between semantic elements from both point cloud and image data is formulated. Subsequently, optimal extrinsic parameters are estimated by minimizing the cost function. We evaluate the proposed method either with GT or predicted semantics on KITTI dataset. Experimental results and comparisons with the baseline method verify the feasibility of the initialization strategy and the accuracy of the calibration approach. In addition, we release the source code at https://github.com/--/SOIC.