Degeneracy in Self-Calibration Revisited and a Deep Learning Solution for Uncalibrated SLAM
This addresses a practical issue for SLAM systems in driving scenarios, offering an incremental improvement over previous geometric and learning approaches.
The paper tackles the problem of self-calibrating camera intrinsics and radial distortion in SLAM, revealing a theoretical ambiguity between radial distortion and scene depth, and proposes a CNN-based solution that achieves comparable or superior performance to existing methods.
Self-calibration of camera intrinsics and radial distortion has a long history of research in the computer vision community. However, it remains rare to see real applications of such techniques to modern Simultaneous Localization And Mapping (SLAM) systems, especially in driving scenarios. In this paper, we revisit the geometric approach to this problem, and provide a theoretical proof that explicitly shows the ambiguity between radial distortion and scene depth when two-view geometry is used to self-calibrate the radial distortion. In view of such geometric degeneracy, we propose a learning approach that trains a convolutional neural network (CNN) on a large amount of synthetic data. We demonstrate the utility of our proposed method by applying it as a checkerboard-free calibration tool for SLAM, achieving comparable or superior performance to previous learning and hand-crafted methods.