PubAg

Main content area

Detecting Bakanae disease in rice seedlings by machine vision

Author:
Chung, Chia-Lin, Huang, Kai-Jyun, Chen, Szu-Yu, Lai, Ming-Hsing, Chen, Yu-Chia, Kuo, Yan-Fu
Source:
Computers and electronics in agriculture 2016 v.121 pp. 404-411
ISSN:
0168-1699
Subject:
Fusarium fujikuroi, Oryza sativa, color, computer vision, conidia, cultivars, grain yield, inflorescences, pathogens, rice, scanners, seed-borne diseases, seedlings, seeds, support vector machines
Abstract:
Bakanae disease, or “foolish seedling”, is a seed-borne disease of rice (Oryza sativa L.). Infected plants can yield empty panicles or perish, resulting in a loss of grain yield. The disease occurs most frequently when contaminated seeds are used. Once the seeds are contaminated, the pathogen Fusarium fujikuroi spreads in the field. Therefore, infected plants must be screened at early developmental stages. This work proposes an approach to nondestructively distinguish infected and healthy seedlings at the age of 3weeks using machine vision. Seeds of the rice cultivars Tainan 11 and Toyonishiki were inoculated with a conidial suspension of F. fujikuroi. The seedling were cultivated in an incubator for 3weeks. The images of infected and control seedlings were acquired using flatbed scanners to quantify their morphological and color traits. Support vector machine (SVM) classifiers were developed for distinguishing the infected and healthy seedlings. A genetic algorithm was used for selecting essential traits and optimal model parameters for the SVM classifiers. The proposed approach distinguished infected and healthy seedlings with an accuracy of 87.9% and a positive predictive value of 91.8%.
Agid:
5255221