In this paper, we present Calibration Recurrent Convolutional Neural Network (CalibRCNN) to infer a 6 degrees of freedom (DOF) rigid body transformation between 3D LiDAR and 2D camera. Different from the existing methods, our 3D-2D CalibRCNN not only uses the LSTM network to extract the temporal features between 3D point clouds and RGB images of consecutive frames, but also uses the geometric loss and photometric loss obtained by the interframe constraint to refine the calibration accuracy of the predicted transformation parameters. The CalibRCNN aims at inferring the correspondence between projected depth image and RGB image to learn the underlying geometry of 2D-3D calibration. Thus, the proposed calibration model achieves a good generalization ability to adapt to unknown initial calibration error ranges, and other 3D LiDAR and 2D camera pairs with different intrinsic parameters from the training dataset. Extensive experiments have demonstrated that our CalibRCNN can achieve stateof-the-art accuracy by comparison with other CNN based methods.
The external calibration between 3D LiDAR and 2D camera is an extremely important step towards multimodal fusion for robot perception. However, its accuracy is still unsatisfactory. To improve the accuracy of calibration, we first analyze the interference factors that affect the performance of the calibration model under a causal inference framework in this study. Guided by the causality analysis, we present Iter-CalibNet (Iterative Calibration Convolutional Neural Network) to infer a 6 degrees of freedom (DoF) rigid body transformation between 3D LiDAR and 2D camera. By downscaling point clouds to obtain more overlapping region between 3D–2D data pair and applying iterative calibration manner, the interference of confounding bias in the calibration model is effectively eliminated. Moreover, our Iter-CalibNet adds non-local neural network after each convolution operation to capture the transformation relationship. We also combine the geometric loss and photometric loss obtained from the interframe constraints to optimize the calibration accuracy. Extensive experiments demonstrate that our Iter-CalibNet can achieve leading performance by comparison with other CNN based and traditional calibration methods.
scite is a Brooklyn-based organization that helps researchers better discover and understand research articles through Smart Citations–citations that display the context of the citation and describe whether the article provides supporting or contrasting evidence. scite is used by students and researchers from around the world and is funded in part by the National Science Foundation and the National Institute on Drug Abuse of the National Institutes of Health.
hi@scite.ai
10624 S. Eastern Ave., Ste. A-614
Henderson, NV 89052, USA
Copyright © 2024 scite LLC. All rights reserved.
Made with 💙 for researchers
Part of the Research Solutions Family.