Classification of vitreousness in durum wheat using soft X-rays and transmitted light images

S. Neethirajan, C. Karunakaran, S. Symons, D.S. Jayas*

*Corresponding author for this work

Research output: Contribution to journalArticleAcademicpeer-review

39 Citations (Scopus)

Abstract

Hardness is a kernel characteristic that influences both milling and processing characteristics of wheat. It is one characteristic that is used for segregating wheat to meet the needs for various products. Kernel virteousness is a visual marker for hardness and is the characteristic assessed during the grading process. The potential of classifying vitreous and non-vitreous durum wheat kernels, in crease-down position, using imaging systems based on real time soft X-rays or transmitted light was assessed in this study. Durum wheat kernels at 14, 15 and 16% moisture contents were used as samples in this study. Image features modeling gray level distribution, textural and shape moments were measured and used to develop a classification system for vitreous and non-vitreous durum wheat kernels. The classification accuracies were 76% for vitreous kernels and 82% for non-vitreous kernels at 16% moisture content using the soft X-ray system but for the transmitted light system, the classification accuracies were 86% for vitreous and 93% for non-vitreous kernels. Moisture content had no effect on classifying vitreous or non-vitreous kernels by the transmitted light system but the classification accuracies increased with moisture content for non-vitreous kernels by the soft X-ray system.

Original languageEnglish
Pages (from-to)71-78
Number of pages8
JournalComputers and Electronics in Agriculture
Volume53
Issue number1
DOIs
Publication statusPublished - Aug 2006
Externally publishedYes

Keywords

  • Image analysis
  • Non-vitreous kernels
  • Soft X-ray images
  • Transmitted light images
  • Vitreous kernels

Fingerprint

Dive into the research topics of 'Classification of vitreousness in durum wheat using soft X-rays and transmitted light images'. Together they form a unique fingerprint.

Cite this