Detection of tungro disease in rice leaf in relation to nitrogen level using laser light backscattering imaging

• Main Author: Bachik, Nur Azizah
• Format: Thesis
• Language: English
• Published: 2017
• Online Access: http://psasir.upm.edu.my/id/eprint/71189/1/FK%202017%2059%20-%20IR.pdf; http://psasir.upm.edu.my/id/eprint/71189/
• Institution: Universiti Putra Malaysia

Rice tungro disease (RTD) is one of the dangerous rice diseases that occurred due to two viruses i.e. rice tungro disease spherical virus (RTSV) and rice tungro disease baciliform virus (RTBV) which is transmitted by green leafhoppers (GLH). Hence, fast detection of this disease is required in order to prevent higher yield loss. The current conventional techniques i.e. serological and nucleic acid techniques are skill dependent and time consuming. Therefore, this study is proposed to investigate the capability of laser light backscattering imaging (LLBI) for detecting RTD in rice leaf. MR-219 rice cultivar was planted in a randomized complete block design (RCDB) with two factorial treatments i.e. inoculation (INO) and nitrogen (N) treatments with three replications. INO treatments including healthy (H), inoculation 1 (INO 1) (3 GLH/seedling) and inoculation 2 (INO 2)(5 GLH/seedling) while N treatments including N1 (0kg/ha), N2 (85 kg/ha), N3 (170 kg/ha) and N4 (250 kg/ha). The seedlings were inoculated with infected GLH during 20 days after planting (DAP). The data collection were carried out on youngest fully expanded leaf at where the readings were taken at three different location i.e. leaf tip, middle leaf and leaf base (4 cm interval). The images then were analyse using image processing toolbox, MatlabR2013a (The Mathworks, Inc. Nattrick, MA, 2013a) to extract LLBI parameters from the images. The polymerase chain reaction (PCR) then was done as a confirmation of the RTD establishment. Results showed that the growth performance i.e. rice plant height and tiller number per hill were influenced by INO treatments but was not influenced by N treatments. Results also indicated significant difference between chlorophyll content and colour parameters i.e. lightness (L*) and chroma (C*) for INO treatments while only colour parameters i.e. red/green (a*) and hue (ho) were significantly influenced by N treatments. The PCR result showed that the expected deoxyribonucleic acid (DNA) size (500 base pairs (bp)) of RTBV was amplified for inoculated rice plants. The LLBI parameters showed that all parameters except maximum intensity, minimum intensity and ratio were influenced by INO treatments while only minimum intensity was influenced by N treatments. There were significant difference between reading points on rice leaf for soil plant analysis development (SPAD) reading, colorimeter parameters and LLBI parameters for INO treatments but no significant for N treatments. The principal component analysis (PCA) and linear discriminant analysis (LDA) showed that LLBI is more appropriate for accessing RTD infection and sensitive in differentiate between reading points location on rice leaf. LDA classification rates for H, INO 1 and INO 2 of LLBI were 83.82%, 75.85% and 78.81%, the colorimeter were 85.60%, 74.46%, and 69.90% while SPAD were 80.95%, 51.19%, and 42.86% respectively.