Structure-From-Motion in 3D Space Using 2D Lidars.

Area of Science:

• Robotics
• Computer Vision
• Geospatial Analysis

Background:

• 2D LiDAR sensors alone offer limited data for 3D pose estimation.
• Existing methods often require multiple sensor types for accurate spatial reconstruction.
• A need exists for efficient structure-from-motion techniques using minimal sensor configurations.

Purpose of the Study:

• To develop a novel structure-from-motion (SfM) methodology utilizing solely 2D LiDAR sensors.
• To enable accurate 6D pose estimation and 3D environmental mapping without supplementary sensors.
• To overcome the inherent limitations of 2D LiDAR data for spatial perception.

Main Methods:

• A novel approach computing system pose from 2D LiDAR scan data, specifically using line segments and corresponding planes.
• Outlier rejection to ensure data integrity before refinement.
• Nonlinear optimization for refining both the estimated 6D pose and the reconstructed 3D structures.

Main Results:

• Demonstrated accuracy in estimating 6D pose using only 2D LiDAR.
• Successful reconstruction of surrounding 3D structures from 2D LiDAR data.
• Validation of the method's robustness through experiments with both synthetic and real-world data.

Conclusions:

• The proposed method effectively performs structure-from-motion using only 2D LiDAR sensors.
• The technique provides an accurate and robust solution for pose estimation and 3D mapping in resource-constrained scenarios.
• This approach offers a viable alternative to multi-sensor systems for specific applications.