Presentation + Paper
20 June 2024 Digital twin for a task-driven multispectral camera design
Wenzhi Liao Sr., Gerardo José Mora Jimena, Lukas De Greve, Abdellatif Bey-Temsamani, Bart De Ketelaere Sr.
Author Affiliations +
Abstract
Spectral imaging allows seeing subtle optical signatures or light wavelengths invisible for human vision. This has potential for numerous industrial applications and a large economic impact by improving quality inspection, increasing automation and the development of innovative applications. However, current commercial spectral imaging systems (both hyper- and multi- spectral) are either too expensive or inflexible to be used in industrial applications (especially for SMEs: small and medium-sized enterprises), whereas standard RGB cameras are often difficult and mostly impractical to classify those materials that are looking visually similar (e.g., real and fake leaves with similar color, shape, texture). In this paper, we propose a new framework for a multispectral camera design. In particular, we first build a digital twin to simulate the output of a camera given a hardware configuration and details on the scene/object, then we optimize the wavelengths according to a specific application defined by end-users for the best filters selections. By combining the filters, illumination, sensor, lenses, pixel size, etc. from digital twin and optimization, we finally design a multispectral camera for the end-users, which provides a low-cost, flexible solution for industrial users. Experimental results from material classification show that our multispectral camera with 3 spectral bands can achieve similar performances as hyperspectral camera on real leaves separation.
Conference Presentation
(2024) Published by SPIE. Downloading of the abstract is permitted for personal use only.
Wenzhi Liao Sr., Gerardo José Mora Jimena, Lukas De Greve, Abdellatif Bey-Temsamani, and Bart De Ketelaere Sr. "Digital twin for a task-driven multispectral camera design", Proc. SPIE 12999, Optical Sensing and Detection VIII, 129991C (20 June 2024); https://doi.org/10.1117/12.3022296
Advertisement
Advertisement
RIGHTS & PERMISSIONS
Get copyright permission  Get copyright permission on Copyright Marketplace
KEYWORDS
Cameras

Tunable filters

Hyperspectral imaging

Optical filters

Image filtering

Digital imaging

Design

Back to Top