- You are here:
-
-
All publications
- Evaluation of segmentation methods for discriminating healthy and fusarium infected corn grains using RGB images
All publications from your selection
Evaluation of segmentation methods for discriminating healthy and fusarium infected corn grains using RGB images
Authors: Rathna Priya Thangaraj Sundaramurthy, Manickavasagan Annamalai
Identifier: CSBE21642
Download file:
Published in: CSBE-SCGAB Technical Conferences » 5th CIGR and AGM Quebec City 2021 » Regular Sessions
Download RAW file:
DOI:
Description: Corn is one of the largest cultivated grain crops and the most predominant crop in Canada. The corn grains in the fields are easily susceptible to fungal infection by Fusarium spp causing Fusarium head blight (FHB) which eventually leads to secondary metabolite production, especially the deoxynivalenol (DON) mycotoxin. FHB creates losses to the farmers and the economy at large. The existing techniques to identify the fusarium infected kernels are expensive and time consuming. Bioimaging techniques have potential for the classification of healthy and infected corn kernels and for implementation in the supply chain for automated cleaning system. In this study, the most common segmentation methods such as graph cut, watershed, clustering and Otsu?s thresholding were evaluated for their ability to clearly segment both the healthy and infected corn kernels. The dataset contained 2-dimensional RGB images of both the germ side and endosperm side of healthy and Fusarium infected corn grains. The images consisted of the grains arranged in a specific line pattern and the grains mixed without any specific arrangement. Jaccard similarity index and Dice coefficients were used to evaluate the segmentation efficiency. The watershed segmentation algorithm performed better for both the healthy and infected kernels and had a performance accuracy of 88%. The classification model developed based on the dataset collected using watershed segmentation algorithm yielded accuracies of 94% and 80% for the training and test set, respectively
Keywords:
Résumé:
Mots-clés:
Citation:
Volume:
Issue:
Pages -
Contributor:
Date: 2021-06-11
Technical field:
Conference name: 5th CIGR International Conference and CSBE-SCGAB AGM 2021, Quebec City,QC, 11-14 May 2021.
Session name: Monitoring, Control and Data Analysis 1
Other information:
Type: Text.Abstract
Format:
Publication type: Presentation
Source:
Relation:
Coverage: Canada
Language 1: en
Language 2:
Rights: Canadian Society for Bioengineering
Notes:
Other files:
DOI:
Description: Corn is one of the largest cultivated grain crops and the most predominant crop in Canada. The corn grains in the fields are easily susceptible to fungal infection by Fusarium spp causing Fusarium head blight (FHB) which eventually leads to secondary metabolite production, especially the deoxynivalenol (DON) mycotoxin. FHB creates losses to the farmers and the economy at large. The existing techniques to identify the fusarium infected kernels are expensive and time consuming. Bioimaging techniques have potential for the classification of healthy and infected corn kernels and for implementation in the supply chain for automated cleaning system. In this study, the most common segmentation methods such as graph cut, watershed, clustering and Otsu?s thresholding were evaluated for their ability to clearly segment both the healthy and infected corn kernels. The dataset contained 2-dimensional RGB images of both the germ side and endosperm side of healthy and Fusarium infected corn grains. The images consisted of the grains arranged in a specific line pattern and the grains mixed without any specific arrangement. Jaccard similarity index and Dice coefficients were used to evaluate the segmentation efficiency. The watershed segmentation algorithm performed better for both the healthy and infected kernels and had a performance accuracy of 88%. The classification model developed based on the dataset collected using watershed segmentation algorithm yielded accuracies of 94% and 80% for the training and test set, respectively
Keywords:
Résumé:
Mots-clés:
Citation:
Volume:
Issue:
Pages -
Contributor:
Date: 2021-06-11
Technical field:
Conference name: 5th CIGR International Conference and CSBE-SCGAB AGM 2021, Quebec City,QC, 11-14 May 2021.
Session name: Monitoring, Control and Data Analysis 1
Other information:
Type: Text.Abstract
Format:
Publication type: Presentation
Source:
Relation:
Coverage: Canada
Language 1: en
Language 2:
Rights: Canadian Society for Bioengineering
Notes:
Other files:
©
2024
CSBE-SCGAB