Early detection of Sclerotinia sclerotiorum on oilseed rape leaves based on optical properties

Abstract

Oilseed rape (Brassica napus L.) is susceptible to Sclerotinia sclerotiorum and its yield would reduce dramatically by the sclerotinia disease. Early detection of the pathogen to prevent the spread of sclerotinia is of great significance. In this paper, the optical response of oilseed rape leaves to Sclerotinia sclerotiorum was analyzed and a new method for early detection of sclerotinia disease based on optical properties was proposed. The optical absorption (μa) and reduced scattering (μs′) coefficients of healthy and infected (invisible-symptom and visible-symptom) oilseed rape leaves were measured by using a single integrating sphere (SIS) system over 500–1000 nm. Results showed that with the development of Sclerotinia sclerotiorum infection from healthy to visible-symptom leaves, μa decreased in 600–700 nm that contained obvious absorption peaks by pigments, while μs′ increased in 500–1000 nm. Then linear discriminant analysis (LDA) and support vector machines (SVM) models were developed to discriminate infected leaves from healthy ones, with the raw data of μa in 500–700 nm, μs′ in 500–700 nm, and (μa-μs′) in 600–700 nm, as well as corresponding effective wavelengths optimally selected by the successive projections algorithm (SPA). Results showed that the LDA models with the raw and SPA-selected data of μa in 500–700 nm, and the SPA-selected (μa-μs′) data, and the SVM model with SPA-selected μa data all provided 100.00% classification accuracy. Overall, this study proved that the optical properties of oilseed rape responded to Sclerotinia sclerotiorum at early stage, and could be a new basis for early detection of sclerotinia disease.

Plant phenotyping relevance

光学特性と機械学習を用いて油糠菜葉の感染状態・病害を早期検出する手法を開発しており、植物の病徴状態の取得・判別が研究の中心である。

Code and data availability

公開本文の所在を確認できませんでした。非公開または購読が必要な可能性があります。

No evidence-backed public reproduction asset is currently recorded.