Development of optical imaging technologies to rapidly assess safety and quality of cereals

Project Information

CEREALSCAN

Grant agreement ID: 330892

Closed project

Start date 1 May 2013

End date 30 April 2016

Funded under

• Specific programme "People" implementing the Seventh Framework Programme of the European Community for research, technological development and demonstration activities (2007 to 2013)

Total cost € 365 124,30

EU contribution € 365 124,30

Coordinated by

UNIVERSITY COLLEGE DUBLIN, NATIONAL UNIVERSITY OF IRELAND, DUBLIN
Ireland

Final Report Summary - CEREALSCAN (Development of optical imaging technologies to rapidly assess safety and quality of cereals)

Project objectives
The overall goal of this project is to develop and validate on-line, non-destructive optical imaging technologies to rapidly assess safety and quality of cereals at critical processing stages post-harvest. This will reduce food safety risks and result in economic benefit to the cereal industry.
In order to overcome the limitations stated above, this research will focus on the following specific objectives:
Objective 1: Development of combined Vis / NIR reflectance, fluorescence and Raman hyperspectral imaging technologies for cereal safety and quality control.
Objective 2: Development of super-resolution algorithms to achieve sub-pixel detection of contaminants at trace levels.
Objective 3: Validation of techniques developed in objectives 1 & 2 on EU indigenous cereals

Work Performed
WP1: Training in macro fluorescence and Raman imaging, (01/05/2013-31/10/2013).
Training on macro fluorescence imaging and Raman imaging was provided at the USDA-BARC using pepper and wheat samples.
WP2: Contaminants detection, damage and quality factors assessment (01/08/2013-30/04/2014).
Set-up of NIR and fluorescence imaging system for imaging acquisition of cereals and seeds has been completed. Multivariate algorithms for image analysis have been developed and optimized
Summary: The NIR and macro fluorescence imaging systems were set-up for analysis of cereal grains and seeds. Dr. Esquerre designed an illuminant diffuser to reduce the effect specular reflection due cereal grains morphology on NIR hyperspectral images. Algorithms for image segmentation, contaminants detection, damage and quality factors assessment has been developed and implemented by Dr. Esquerre in MATLAB for each of the imaging techniques.
WP3: Data fusion of Vis/NIR, fluorescence and Raman spectral imaging (01/05/2014-30/01/2015)
Data fusion algorithms were developed and evaluated to exploit the synergy of visible / NIR reflectance, fluorescence and Raman imaging for the safety and quality parameters of wheat samples. Three levels of data fusion were investigated:
a) Low level: obtained by overlapping the hypercubes or images corresponding to the selected wavelengths of the three spectral technologies.
b) Medium level: obtained by superimposing the relevant features of each source.
c) High level: obtained by fusing the values predicted with the models developed individually with each one to the spectral techniques used

WP4: Super-resolution on optimal models (01/11/2014-31/07/2015)
Suitability of algorithms for multi-frame image super-resolution were tested using data acquired with the optical spectral systems at BARC. These algorithms gain additional information from the sub-pixel spatial shift in the multiple images of the sample. The increased resolution was investigated for improvement of accuracy and limit of detection on models developed.
WP5 & WP6: Validation on European indigenous cereals (01/05/2015-30/04/2016).
Validation of best models developed in WP2-4 for assessment the extent of damage by waxiness and quality in European indigenous wheat and barley.

Main Results Achieved
WP1: Training in macro fluorescence and Raman imaging, (01/05/2013-31/10/2013).
Training stage completed.
WP2: Contaminants detection, damage and quality factors assessment (01/08/2013-30/04/2014).
Set-up of NIR and fluorescence imaging system for imaging acquisition of cereals and seeds has been completed. Algorithms for contaminants detection, damage and quality factors assessment has been developed..
WP3: Data fusion of Vis/NIR, fluorescence and Raman spectral imaging (01/05/2014-30/01/2015)
Data fusion algorithms were developed and evaluated.
WP4: Super resolution on optimal models (01/11/2014-31/07/2015)
Algorithms for multi-frame image super-resolution are under development and evaluation.
WP5 & WP6: Validation on European indigenous cereals (01/05/2015-30/04/2016).

Expected final results, their potential impact and use
The research identified synergy between NIR, fluorescence and Raman imaging to rapidly assess safety and quality of cereals at critical processing stages post-harvest. This work contributes to the development of on-line/at-line, non-destructive optical imaging technologies that will reduce food safety risks and result in economic benefit to the cereal industry.