Hyperspectral image reconstruction using multi-colour and time-multiplexed LED illumination

Julius Tschannerl, Jinchang Ren, Huimin Zhao, Fu-Jen kao, Stephen Marshall, Peter Yuen

Research output: Contribution to journalArticle

Abstract

The rapidly rising industrial interest in hyperspectral imaging (HSI) has generated an increased demand for cost effective HSI devices. We are demonstrating a mobile and low-cost multispectral imaging system, enabled by time-multiplexed RGB Light Emitting Diodes (LED) illumination, which operates at video framerate. Critically, a deep Multi-Layer Perceptron (MLP) with HSI prior in the spectral range of 400–950 nm is trained to reconstruct HSI data. We incorporate regularisation and dropout to compensate for overfitting in the largely ill-posed problem of reconstructing the HSI data. The MLP is characterised by a relatively simple design and low computational burden. Experimental results on 51 objects of various references and naturally occurring materials show the effectiveness of this approach in terms of reconstruction error and classification accuracy. We were also able to show that utilising additional colour channels to the three R, G and B channels adds increased value to the reconstruction.
Original languageEnglish
Pages (from-to)352-357
Number of pages6
JournalOptics and Lasers in Engineering
Volume121
Early online date6 May 2019
DOIs
Publication statusPublished - 31 Oct 2019

Fingerprint

self organizing systems
image reconstruction
Image reconstruction
Light emitting diodes
light emitting diodes
Lighting
illumination
Color
dropouts
color
Multilayer neural networks
costs
Imaging systems
Costs
Hyperspectral imaging

Keywords

  • hyperspectral Imaging
  • deep learning
  • spectral reconstruction
  • LED illumination

Cite this

@article{8b0b2e53dc2d4b50ac2d3e7d00e71dfd,
title = "Hyperspectral image reconstruction using multi-colour and time-multiplexed LED illumination",
abstract = "The rapidly rising industrial interest in hyperspectral imaging (HSI) has generated an increased demand for cost effective HSI devices. We are demonstrating a mobile and low-cost multispectral imaging system, enabled by time-multiplexed RGB Light Emitting Diodes (LED) illumination, which operates at video framerate. Critically, a deep Multi-Layer Perceptron (MLP) with HSI prior in the spectral range of 400–950 nm is trained to reconstruct HSI data. We incorporate regularisation and dropout to compensate for overfitting in the largely ill-posed problem of reconstructing the HSI data. The MLP is characterised by a relatively simple design and low computational burden. Experimental results on 51 objects of various references and naturally occurring materials show the effectiveness of this approach in terms of reconstruction error and classification accuracy. We were also able to show that utilising additional colour channels to the three R, G and B channels adds increased value to the reconstruction.",
keywords = "hyperspectral Imaging, deep learning, spectral reconstruction, LED illumination",
author = "Julius Tschannerl and Jinchang Ren and Huimin Zhao and Fu-Jen kao and Stephen Marshall and Peter Yuen",
year = "2019",
month = "10",
day = "31",
doi = "10.1016/j.optlaseng.2019.04.014",
language = "English",
volume = "121",
pages = "352--357",
journal = "Optics and Lasers in Engineering",
issn = "0143-8166",

}

Hyperspectral image reconstruction using multi-colour and time-multiplexed LED illumination. / Tschannerl, Julius; Ren, Jinchang; Zhao, Huimin; kao, Fu-Jen; Marshall, Stephen; Yuen, Peter.

In: Optics and Lasers in Engineering, Vol. 121, 31.10.2019, p. 352-357.

Research output: Contribution to journalArticle

TY - JOUR

T1 - Hyperspectral image reconstruction using multi-colour and time-multiplexed LED illumination

AU - Tschannerl, Julius

AU - Ren, Jinchang

AU - Zhao, Huimin

AU - kao, Fu-Jen

AU - Marshall, Stephen

AU - Yuen, Peter

PY - 2019/10/31

Y1 - 2019/10/31

N2 - The rapidly rising industrial interest in hyperspectral imaging (HSI) has generated an increased demand for cost effective HSI devices. We are demonstrating a mobile and low-cost multispectral imaging system, enabled by time-multiplexed RGB Light Emitting Diodes (LED) illumination, which operates at video framerate. Critically, a deep Multi-Layer Perceptron (MLP) with HSI prior in the spectral range of 400–950 nm is trained to reconstruct HSI data. We incorporate regularisation and dropout to compensate for overfitting in the largely ill-posed problem of reconstructing the HSI data. The MLP is characterised by a relatively simple design and low computational burden. Experimental results on 51 objects of various references and naturally occurring materials show the effectiveness of this approach in terms of reconstruction error and classification accuracy. We were also able to show that utilising additional colour channels to the three R, G and B channels adds increased value to the reconstruction.

AB - The rapidly rising industrial interest in hyperspectral imaging (HSI) has generated an increased demand for cost effective HSI devices. We are demonstrating a mobile and low-cost multispectral imaging system, enabled by time-multiplexed RGB Light Emitting Diodes (LED) illumination, which operates at video framerate. Critically, a deep Multi-Layer Perceptron (MLP) with HSI prior in the spectral range of 400–950 nm is trained to reconstruct HSI data. We incorporate regularisation and dropout to compensate for overfitting in the largely ill-posed problem of reconstructing the HSI data. The MLP is characterised by a relatively simple design and low computational burden. Experimental results on 51 objects of various references and naturally occurring materials show the effectiveness of this approach in terms of reconstruction error and classification accuracy. We were also able to show that utilising additional colour channels to the three R, G and B channels adds increased value to the reconstruction.

KW - hyperspectral Imaging

KW - deep learning

KW - spectral reconstruction

KW - LED illumination

U2 - 10.1016/j.optlaseng.2019.04.014

DO - 10.1016/j.optlaseng.2019.04.014

M3 - Article

VL - 121

SP - 352

EP - 357

JO - Optics and Lasers in Engineering

JF - Optics and Lasers in Engineering

SN - 0143-8166

ER -