Mgr. Jaroslav Moravec

Cameras are the prevalent sensors used for perception in autonomous robotic systems, but their initial calibration may degrade over time due to dynamic factors. This may lead to a failure of downstream tasks, such as simultaneous localization and mapping (SLAM) or object recognition. Hence, a computationally lightweight process that detects the decalibration is of interest. We describe a modification of StOCaMo, an online calibration monitoring procedure for a stereoscopic system. The method uses robust kernel correlation based on epipolar constraints; it validates extrinsic calibration parameters on a single frame with no temporal tracking. In this paper, we present a modified StOCaMo with an improved recall rate on small decalibrations through a confirmation technique based on resampled variance. With fixed parameters learned on a realistic synthetic dataset from CARLA, StOCaMo and its proposed modification were tested on multiple sequences from two real-world datasets: KITTI and EuRoC MAV. The modification improved the recall of StOCaMo by 25 % (to 91 % and 82 %, respectively), and the accuracy by 12 % (to 94.7 % and 87.5 %, respectively), while labeling at most one-third of the input data as uninformative. The upgraded method achieved the rank correlation between StOCaMo V-index and downstream SLAM error of 0.78 (Spearman).

The relative poses of visual perception sensors distributed over a vehicle's body may vary due to dynamic forces, thermal dilations, or minor accidents. This paper proposes two methods, OCAMO and LTO, that monitor and track the LiDAR-Camera extrinsic calibration parameters online. Calibration monitoring provides a certificate for reference calibration parameters validity. Tracking follows the calibration parameters drift in time. OCAMO is based on an adaptive online stochastic optimization with a memory of past evolution. LTO uses a fixed-grid search for the optimal parameters per frame and without memory. Both methods use low-level point-like features and a robust kernel-based loss function and work with a small memory footprint and computational overhead. Both include a preselection of informative data that limits their divergence. The statistical accuracy of both calibration monitoring methods is over 98%, whereas OCAMO monitoring can detect small decalibrations better, and LTO monitoring reacts faster on abrupt decalibrations. The tracking variants of both methods follow random calibration drift with an accuracy of about 0.03° in the yaw angle.

Cameras are the prevalent sensors used for perception in autonomous robotic systems, but initial calibration may degrade over time due to dynamic factors. This may lead to the failure of the downstream tasks, such as simultaneous localization and mapping (SLAM) or object recognition. Hence, a computationally light process that detects the decalibration is of interest. We propose StOCaMo, an online calibration monitoring procedure for a stereoscopic system. StOCaMo is based on epipolar constraints; it validates calibration parameters on a single frame with no temporal tracking. The main contribution is the use of robust kernel correlation, which is shown to be more effective than the standard epipolar error. StOCaMo was tested on two real-world datasets: EuRoC MAV and KITTI. With fixed parameters learned on a realistic synthetic dataset from CARLA, it achieved 96.2% accuracy in decalibration detection on EuRoC and KITTI. In the downstream task of detecting SLAM failure, StOCaMo achieved 87.3% accuracy, and its output has a rank correlation of 0.77 with the SLAM error. These results outperform a recent method by Zhong et al., 2021.