Image analysis Developing of prototype model for anthracnose detection in mature green stage of mango fruit with image analysis technique
Research Title : Image analysis Developing of prototype model for anthracnose detection in mature green stage of mango fruit with image analysis technique
Researcher : Dr.Thitima Wongsheree
Office : King Mongkut’s University of Technology Thonburi
Research Grants : Public and Private Linkage for Commercialize Research Project
Year : 2011
Abstract
This study on nondestructive detection of anthracnose infected in Nam Dork Mai mangoes using image processing and near infrared spectroscopy (NIRs) techniques. Mature green Nam Dork Mai mangoes were harvested from exporting mango orchards in Chacherngsao and Ang Thong provinces and immediately transported to the postharvest technology laboratory, KMUTT, Bangkhuntein campus by refrigerator truck. Cleaned fruits and divided for experiments in 2 groups; one group inoculated mango fruits with 106 spore/ml anthracnose spore suspension (inoc) and the other group inoculated with de-water as the control treatment. Prepare the fruit by drawing the rectangle area of 5×4 centimeter in the middle of peel fruit and inoculated with of 400 µL of 106 spore/ml anthracnose spore suspension. The photograph of inoculated area were collected every 05-1.0 hr until anthracnose symptom appearance. The edge area and color image were processing with entropy and image processing technique. From the result, image analysis techniques with entropy processing of edge area gave the linear model. This model could classify anthracnose infection in the 4th hour after inoculation. However, color image processing result of the pixel number was not sufficient for predicted anthracnose infection in mango peel. Our studied showed that image processing with entropy processing have potential to classify the infected mango group. However, this technique and color image was inappropriate technique for predicting natural infected.
Keywords : mango, Anthracnose, NIR