Grant: Strategic Projects - Group
University of Manitoba
$99,500.00
Integration of near-infrared spectroscopy and machine vision for quality assessment of cereal grains
FY 2006-07
Summary
INTRODUCTION: Real-time inspection of products is already in commercial use in automotive, electronics, and other industries. Application of technologies such as machine vision to agricultural objects is more complex because these objects vary in size, shape, colour, and texture. Also for these objects, physical appearance alone does not give a characterization of their usefulness. For example in a lot of wheat, only a few sprouted kernels which may appear sound can have a significant detrimental effect on bread baking quality. Therefore, there is a need to develop systems that can assess the external features of grain along with intrinsic quality characteristics. We are proposing to integrate machine vision and near-infrared (NIR) spectroscopy based systems to determine external grading factors and to predict end-use functionality of wheat. The developed technology can be applied to other cereals and oilseeds with further research. SCIENTIFIC APPROACH: Visible and NIR images along with NIR spectral data of samples of wheat of different grades, grown in three different years and at 20 different locations in Western Canada, will be acquired. Algorithms will be developed to extract morphological, colour, and textural features of the kernels presented individually and in bulk. Our algorithms for separating touching objects will be developed further using various theoretical approaches and made speedier to execute. Algorithms will also be developed to determine the features of kernels which will be partially exposed in a bulk sample requiring fundamental research in machine vision and pattern recognition. Spectral data will analyzed by implementing wavelet transform and artificial neural networks. Mechanical systems will be developed to present samples in a non-touching manner and in bulk for scanning the samples. EXPECTED SIGNIFICANCE: The integrated system will have a direct impact on the Canadian grain industry. The results of this research will help Canada in becoming more globally competitive and thus result in increased financial gain to grain industry and farmers.
Program
Strategic Projects - Group
Recipient
Location
Manitoba, Canada
Recipient Type
Academia
Type
Grant
Source
NSERC's Awards Data
As published
Application id
322266-2005
Name nom
Jayas, Digvir
Department département
Biosystems Engineering
Organization id
19
Institution établissement
University of Manitoba
Province en
Manitoba
Province fr
Manitoba
Country en
CANADA
Country fr
CANADA
Fiscal year exercice financier
2006
Competition year année de concours
2005
Award amount
99500
Program id
STPGP
Program name en
Strategic Projects - Group
Program name fr
Projets stratégiques - De groupe
Group en
Research partnerships programs
Group fr
Programmes de partenariat
Committee code
1002
Committee name en
Strategic Projects Selection Panel - Biosciences B
Committee name fr
Comité de sélection des projets stratégiques - biosciences B
Area of application code
100
Area of application group en
Agriculture and primary food production
Area of application group fr
Agriculture et produits alimentaires primaires
Area of application en
Agriculture and primary food production
Area of application fr
Agriculture et produits alimentaires primaires
Research subject code
1200
Research subject group en
Agricultural engineering
Research subject group fr
Génie rural
Research subject en
Agricultural engineering
Research subject fr
Génie rural
Partie
2
Nb partie
3
Application title
Integration of near-infrared spectroscopy and machine vision for quality assessment of cereal grains
Keyword
NA
Application summary
INTRODUCTION: Real-time inspection of products is already in commercial use in automotive, electronics, and other industries. Application of technologies such as machine vision to agricultural objects is more complex because these objects vary in size, shape, colour, and texture. Also for these objects, physical appearance alone does not give a characterization of their usefulness. For example in a lot of wheat, only a few sprouted kernels which may appear sound can have a significant detrimental effect on bread baking quality. Therefore, there is a need to develop systems that can assess the external features of grain along with intrinsic quality characteristics. We are proposing to integrate machine vision and near-infrared (NIR) spectroscopy based systems to determine external grading factors and to predict end-use functionality of wheat. The developed technology can be applied to other cereals and oilseeds with further research. SCIENTIFIC APPROACH: Visible and NIR images along with NIR spectral data of samples of wheat of different grades, grown in three different years and at 20 different locations in Western Canada, will be acquired. Algorithms will be developed to extract morphological, colour, and textural features of the kernels presented individually and in bulk. Our algorithms for separating touching objects will be developed further using various theoretical approaches and made speedier to execute. Algorithms will also be developed to determine the features of kernels which will be partially exposed in a bulk sample requiring fundamental research in machine vision and pattern recognition. Spectral data will analyzed by implementing wavelet transform and artificial neural networks. Mechanical systems will be developed to present samples in a non-touching manner and in bulk for scanning the samples. EXPECTED SIGNIFICANCE: The integrated system will have a direct impact on the Canadian grain industry. The results of this research will help Canada in becoming more globally competitive and thus result in increased financial gain to grain industry and farmers.
