Related Experiment Video
Updated: Jul 4, 2025
07:05
Applying Hyperspectral Reflectance Imaging to Investigate the Palettes and the Techniques of Painters
Published on: June 18, 2021
2.4K
Spectral reflectance reconstruction based on wideband multi-illuminant imaging and a modified particle swarm
Optics Express
|February 1, 2024
Summary
This study introduces a novel method for spectral reflectance reconstruction using multi-illuminant imaging and optimized algorithms. Combining illuminants significantly improves accuracy, offering a simple, effective solution for color measurement.
Area of Science:
- Color Science
- Image Processing
- Computational Optimization
Background:
- Accurate spectral reflectance factor reconstruction is crucial for color reproduction and measurement.
- Traditional methods often face limitations in accuracy and applicability across diverse lighting conditions.
- Wideband multi-illuminant imaging offers potential for enhanced spectral reconstruction.
Purpose of the Study:
- To develop and evaluate a novel method for spectral reflectance factor reconstruction using wideband multi-illuminant imaging.
- To investigate the impact of illuminant combinations, training data, and optimization algorithms on reconstruction accuracy.
- To assess the performance of the proposed method compared to existing algorithms.
Main Methods:
- Utilized a programmable LED lighting system to generate nine light sources with varying correlated color temperatures (1924 K - 15746 K).
- Captured images of three color charts (X-Rite ColorChecker Digital SG, SCOCIE ScoColor paint, SCOCIE ScoColor textile) under the generated illuminants.
- Employed modified Bare Bones Particle Swarm Optimization (BBPSO) algorithms for spectral reconstruction model training and testing.
Main Results:
- Combinations of two illuminants with significantly different correlated color temperatures more than doubled reconstruction accuracy compared to single illuminants.
- Training with colors covering a wider color space region resulted in more accurate spectral reflectance factor reconstructions.
- The proposed method achieved average reconstruction errors of 0.94 and 1.08 CIEDE2000 units for specific color charts under A+D90 illumination.
Conclusions:
- The proposed multi-illuminant imaging method, coupled with modified BBPSO, significantly enhances spectral reflectance factor reconstruction accuracy.
- The approach is simple, requires no prior knowledge, and is readily implementable in non-contact color measurement systems.
- Certain existing algorithms are unsuitable for multi-illuminant spectral reconstruction due to optimization complexity.
