Please use this identifier to cite or link to this item:
Title: 3D Color co-occurrence texture features as tool to evaluate quality of fruits
Authors: Bama, B Sathya
Harinie, T
Chellam, I Janani
Raju, S
Abhaikumar, V
Keywords: 3D Color co-occurrence distribution
Histogram based thresholding
Sum of squared distance (SSD)
Issue Date: Nov-2011
Publisher: NISCAIR-CSIR, India
Abstract: This paper proposes a novel quality evaluation method to categorize fruits using color and texture features. In proposed method, input image is segmented using histogram based thresholding technique to identify whether the portion of image is defective or non defective. For defected samples, texture features are extracted from 3D color co-occurrence distribution. Sum of squared distance (SSD) is calculated between texture features of training and test fruits. Non defective fruits are considered as extra class, whereas classes I and II are moderately and highly defective fruits respectively. A database of 150 fruit samples was collected from mozambi and orange fruits with all quality classes for marketable – grade fruit inspection and distribution system. Images (80) were manually classified for training and the rest were used for testing. Proposed method (classification accuracy,> 93%) outperforms existing methods.
Description: 912-917
ISSN: 0975-1084 (Online); 0022-4456 (Print)
Appears in Collections:JSIR Vol.70(11) [November 2011]

Files in This Item:
File Description SizeFormat 
JSIR 70(11) 912-917.pdf154.1 kBAdobe PDFView/Open

Items in NOPR are protected by copyright, with all rights reserved, unless otherwise indicated.