Discriminant Analysis with Visible Lighting Properties for White Root Disease Infected Rubber Tree

Abstract

The most serious disease known in rubber industry is root disease and among the major root diseases, white root is the most destructive agent of trees and agricultural crops. Since, it is too difficult and expensive to treat root disease infection on trees; prevention is important whereby one must rely on symptoms appearing on roots themselves in order to recognize the disease. Infection symptom on the tree is detected when leaves became discolored yellowish and dying. The main objective of this work is to investigate empirically an infected rubber trees infected by white root disease where its symptom could be detected visually from leaves gradual discolouration. Visible spectrum of optical measurements is taken on four different regions of interest (ROI) locations of the top side leaf sample features such as petiolule, main vein/midrib, vein and leaf cell of rubber trees. Statistical techniques is used to analyse for conclusive scientific findings of which ROIs above has shown clear discrimination between the healthy, medium and worst condition. The scope of work involves raw data inclusion of leaf samples belong only to 2025 rubber tree clone. This clone is recommended by Rubber Research Institute of Malaysia (RRIM) management due to its popularity and commercially used by small scale planters. Outcomes of this work has suggested that only main vein/midrib and leaf cell ROIs produced convincing significant discrimination between healthy, medium and worst case. Thus, their measurements can be recommended for developing on the shelf technology engineering sensor instrument using non-invasive advanced signal processing techniques and intelligent system for early detection of white root disease.