Related Experiment Video
Updated: Oct 2, 2025
08:39
Shaping the Amplitude and Phase of Laser Beams by Using a Phase-only Spatial Light Modulator
Published on: January 28, 2019
9.9K
PAL-SLAM: a feature-based SLAM system for a panoramic annular lens
Optics Express
|February 25, 2022
Summary
This study introduces PAL-SLAM, a new system for robots and self-driving cars that accurately maps surroundings even during fast movements or when GPS is unavailable. It offers precise localization below 1 cm, improving navigation reliability.
Area of Science:
- Robotics and Computer Vision
- Simultaneous Localization and Mapping (SLAM)
Background:
- Traditional visual SLAM systems struggle with rapid motion and loop detection.
- Autonomous systems require robust localization, especially in challenging environments like urban areas.
Purpose of the Study:
- To develop a feature-based Simultaneous Localization and Mapping (SLAM) system tailored for Panoramic-Annular-Lens (PAL) cameras.
- To enhance navigation accuracy and reliability in autonomous driving and robotics.
Main Methods:
- Utilized a mask to extract and match features within the annular effective area of PAL camera images.
- Developed a precise PAL-camera model for transforming features into unit vectors.
- Implemented a novel inlier-checking metric as an epipolar constraint for system initialization.
Main Results:
- PAL-SLAM achieved typical accuracy below 1 cm on large-scale indoor and outdoor datasets (>12,000 images).
- The system demonstrated consistent performance during rapid camera rotations and Global Navigation Satellite System (GNSS) signal blockage.
- PAL-SLAM successfully detected both unidirectional and bidirectional loop closures.
Conclusions:
- PAL-SLAM offers a highly accurate and robust localization solution for autonomous systems using PAL cameras.
- It serves as a viable supplement or alternative to expensive navigation systems, particularly in GNSS-denied urban environments.
