Master of Science in Robotics Thesis Talk

  • Remote Access Enabled - Zoom
  • Virtual Presentation
  • Masters Student
  • Robotics Institute
  • Carnegie Mellon University
Master's Thesis Presentation

On-the-fly Targetless Extrinsics Calibration For Multi-Stereo Systems Without Field-of-View Overlap

In this talk, we propose an on-the-fly extrinsics calibration method for stereo pairs lacking overlapping field of view that is robust to visual odometry errors. Multi-stereo systems are becoming increasingly popular because of their large field of view (FoV) that benefits both state estimation and mapping. A critical challenge in multi-stereo systems is to calibrate the extrinsics among the stereo pairs, which becomes even more difficult when the FoV of different stereo pairs does not overlap. Moreover, due to external forces (e.g., impact or vibration) or changes of the environment (e.g. temperature or pressure), the extrinsics can change over time. As a result, on-the-fly calibration of the extrinsics is necessary. We propose an on-the-fly targetless extrinsic calibration method for a multi-stereo system without FoV overlap.  Experimental results with both simulation and real-world data show that the proposed method is successful in estimating the extrinsics on-the-fly and is fairly robust to outliers and degeneracy in the measurements.

Thesis Committee:
Michael Kaess (Advisor)
George Kantor
Alexander Spitzer

Zoom Participation Enabled. See announcement.

For More Information, Please Contact: