Minimal Solvers for Full-DoF Motion Estimation from Asynchronous Differential SfM
Researchers have developed a new framework for estimating egomotion using asynchronous optical flow from event cameras. This method allows for the recovery of both angular and linear velocities, overcoming challenges posed by the asynchronous data streams of these sensors. The proposed optimization algorithm and a novel algebraic minimal 5-point solver enable full degree of freedom egomotion estimation, outperforming traditional synchronous methods in accuracy and robustness. AI
IMPACT Establishes a foundation for improved continuous-time motion estimation in high-speed robotics.