Classification of tomato (Solanum lycopersicum) leaf diseases using color texture features and Artificial Neural Networks (ANN) .

Date

10-2011

Degree

Bachelor of Science in Applied Physics

College

College of Arts and Sciences (CAS)

Adviser/Committee Chair

Alexandra B. Santos

Co-adviser

Darwin B. Putungan

Abstract

A method for identifying healthy and diseased regions of tomato leaves was developed. This was implemented by performing texture feature analysis on the Gray-Level Co-Occurrence Matrices (GLCM) images of the leaves. Different regions were classified using a trained two-layer feedforward backpropagation artificial neural network. Early Blight regions were detected with up to 90.9% accuracy; Phoma blight regions, 86.7%; and healthy regions, 100%. It was also found that the Hue color layer is better used when differentiating between Early and Phoma Blight, and the Intensity color layer when comparing diseased regions to healthy ones. Keywords: Image Processing (07.05.Pj); Neural networks (07.05.Mh); Matrix Theory (02.10.Yn)

Language

English

Location

UPLB Main Library Special Collections Section (USCS)

Call Number

Thesis

Document Type

Thesis

This document is currently not available here.

Share

COinS