Related Experiment Video
Updated: Jul 4, 2025
15:18
Near Infrared Optical Projection Tomography for Assessments of β-cell Mass Distribution in Diabetes Research
Published on: January 12, 2013
16.4K
Camera optimal projection model identification and calibration method based on the NGO-BA architecture.
Applied Optics
|January 31, 2024
Summary
A new generic camera calibration method unifies diverse lens models for accurate imaging across wide fields of view (FOV). This projection model optimization strategy minimizes reprojection error, enhancing calibration for various camera systems.
Area of Science:
- Computer Vision
- Optical Engineering
- Metrology
Background:
- Diverse camera lens principles and structures necessitate unified imaging models for accurate characterization.
- Existing calibration methods often struggle with a wide range of field-of-view (FOV) and varied projection models.
Purpose of the Study:
- To propose a generic camera calibration method adaptable to various projection models and lens structures.
- To achieve accurate imaging characteristic descriptions across a broad spectrum of FOV.
Main Methods:
- Developed piecewise and polynomial functions for geometric and fitting projection models.
- Implemented a multistation self-calibration bundle adjustment (BA) module for diverse projection models.
- Integrated BA with a northern goshawk optimization architecture for iterative parameter optimization.
Main Results:
- Achieved low reprojection (RP) root mean square errors across various FOVs: 1/20 pixel (68°), 1/13 pixel (84°), 1/9 pixel (115°, 135°), and 1/6 pixel (180°).
- Demonstrated fast and versatile optimization for multiple projection model types and camera systems.
Conclusions:
- The proposed method provides a unified approach for camera calibration, accommodating diverse lens designs and FOVs.
- Efficiently determines optimal projection models and imaging parameters, improving accuracy and applicability in computer vision and optical metrology.
