Related Experiment Video
Updated: Dec 21, 2025
06:08
Automatic Identification of Dendritic Branches and their Orientation
Published on: September 17, 2021
2.2K
Identifying multiple line-structured lights from images via a local-to-global graph representation.
Optics Express
|May 15, 2020
Summary
This study introduces a novel graph framework to efficiently identify multiple line-structured lights for 3D reconstruction. The method resolves ambiguities and achieves high precision without complex laser coding.
Area of Science:
- Computer Vision
- 3D Reconstruction
- Robotics
Background:
- Identifying multiple line-structured lights is crucial for active 3D visual reconstruction.
- Existing methods often rely on complex coding schemes, limiting real-time applications.
Purpose of the Study:
- To develop an efficient algorithm for identifying multiple line-structured lights.
- To overcome the limitations of existing time-consuming and complex coding approaches.
Main Methods:
- A local-to-global graph framework is proposed to model the hierarchical distribution of light segments.
- Lights are grouped into local graphs based on an overlapping metric.
- Local graphs are unified into a global graph using node depth, identifying lights from the same laser plane.
Main Results:
- The algorithm effectively identifies scattered light segments and demonstrates robustness to varying sensor poses.
- Applied to 3D reconstruction, it achieved a precision of 0.025mm.
- The approach eliminates the need for complex auxiliary laser coding.
Conclusions:
- The proposed graph-based method offers a convenient and efficient solution for multi-line light identification in 3D reconstruction.
- It enables real-time sparse 3D reconstruction with high accuracy.
