An edge texture features based methodology for bulk paddy variety recognition

Authors

  • Basavaraj S Anami K L E Institute of Technology, Hubli
  • Naveen Nagendrappa Malvade KLEIT, Hubli
  • N G Hanamaratti University of Agricultural Sciences, Dharwar

Keywords:

paddy, canny, sobel, texture features, feature extraction, ANN, pattern recognition

Abstract

The paper presents a method for recognition of paddy varieties from their bulk grain sample edge images based on Haralick texture features extracted from grey level co-occurrence matrices. The edge images were obtained using Canny and maximum gradient edge detection methods. The average paddy variety recognition performances of the two categories of edge images were evaluated and compared. A feature set of thirteen texture features was considered and the feature set was reduced based on contribution of each feature to the paddy variety recognition accuracy. The average paddy variety recognition accuracy of 87.80% was obtained for the reduced eight texture features extracted from maximum gradient edge images. The work is useful in developing a machine vision system for agriculture produce market and developing multimedia applications in agriculture sciences.

Author Biographies

Basavaraj S Anami, K L E Institute of Technology, Hubli

Professor, Department of Computer Science and Engineering

Naveen Nagendrappa Malvade, KLEIT, Hubli

Information Science and Technology

N G Hanamaratti, University of Agricultural Sciences, Dharwar

Senior Scientist, Department of Genetics and Plant Breeding.

Downloads

Published

2016-03-22

Issue

Vol. 18 No. 1 (2016): CIGR Journal

Section

VII-Information Systems