Reproduce “PL-EVIO: Robust Monocular Event-based Visual Inertial Odometry with Point and Line Features” on OpenResearch
Turn your favorite coding agents into research agents that review literature, develop hypotheses, and run experiments.
Universal app for Apple silicon and Intel. Requires macOS 11 or later.
About this paper
Weipeng Guan, Peiyu Chen, Yuhan Xie, Peng Lu
PL-EVIO introduces a robust monocular visual-inertial odometry system that combines event, standard, and IMU data, uniquely integrating line features from event streams for enhanced performance. The system achieves state-of-the-art accuracy in high-dynamic-range and aggressive motion environments, providing 1000 Hz pose feedback for real-time onboard autonomous quadrotor control.
Key insights
- The Time Surface with polarity (T_p(x,t)) is essential for robust event-corner feature tracking, capturing sufficient gradient information without introducing motion blur.
- Line features extracted from event data offer crucial structural constraints in human-made environments, significantly enhancing robustness and accuracy where point features may be insufficient.
- Tightly fusing event, image, and IMU data allows the system to overcome individual sensor limitations, achieving robust performance across aggressive motion, HDR, and static conditions.
Summary by alphaXiv.
Tip: change arxiv.org to autoarxiv.org in any arXiv abstract link to open the paper here.