High-impact stereo datasets have suboptimal calibration!
We demonstrate the evaluation of the calibration accuracy of stereo cameras. In the video, the evaluation is done on well-known stereo vision datasets (KITTI, Apolloscape, Argoverse, ...).
The approach does not need any calibration pattern.
Additionally, we show the optimization of calibration parameters which leads to smaller errors on these datasets.
The method is designed for online-calibration. It can be implemented in any industrial stereo application to improve the parameters while processing.
--
Detailed information in our paper:
K. Cordes and H. Broszio: "Accuracy Evaluation and Improvement of the Calibration of Stereo Vision Datasets", ECCV - VCAD workshop, 2024
Sept 30, 2024
vcad-workshop.github.io/