Hyperspectral Imaging for Presymptomatic Detection of Tobacco Disease with Successive Projections Algorithm and Machine-learning Classifiers

Hongyan Zhu,Linjun Jiang,Fei Liu,Chu Zhang,Bingquan Chu,Yong He

doi:10.1038/s41598-017-04501-2

Hongyan Zhu, Linjun Jiang + Show 4 more

Open Access

https://doi.org/10.1038/s41598-017-04501-2

Copy DOI

Journal: Scientific Reports	Publication Date: Jun 23, 2017
Citations: 128	License type: open-access

Affiliation: Zhejiang University

Abstract

We investigated the feasibility and potentiality of presymptomatic detection of tobacco disease using hyperspectral imaging, combined with the variable selection method and machine-learning classifiers. Images from healthy and TMV-infected leaves with 2, 4, and 6 days post infection were acquired by a pushbroom hyperspectral reflectance imaging system covering the spectral range of 380–1023 nm. Successive projections algorithm was evaluated for effective wavelengths (EWs) selection. Four texture features, including contrast, correlation, entropy, and homogeneity were extracted according to grey-level co-occurrence matrix (GLCM). Additionally, different machine-learning algorithms were developed and compared to detect and classify disease stages with EWs, texture features and data fusion respectively. The performance of chemometric models with data fusion manifested better results with classification accuracies of calibration and prediction all above 80% than those only using EWs or texture features; the accuracies were up to 95% employing back propagation neural network (BPNN), extreme learning machine (ELM), and least squares support vector machine (LS-SVM) models. Hence, hyperspectral imaging has the potential as a fast and non-invasive method to identify infected leaves in a short period of time (i.e. 48 h) in comparison to the reference images (5 days for visible symptoms of infection, 11 days for typical symptoms).

Highlights

With conventional RGB imaging, near infrared (NIR) spectroscopy and multispectral imaging, HSI integrates conventional imaging and spectroscopy to obtain both spatial and spectral information simultaneously from a sample at spatial resolutions varying from the level of single cells up to macroscopic objects[6, 8, 9]
Extreme learning machine model was developed for the classification of healthy, early blight and late blight of detached leaves, which yielded about 97.1–100% classification accuracy when classifying diseased plants based on spectral information
This outmost goal was achieved by meeting the following specific objectives: (i) determining the corresponding effective wavelengths (EWs) which give the highest correlation between the spectral data and different disease stages; (ii) extracting texture features based on grey-level co-occurrence matrix (GLCM) at the selected EWs; (iii) developing and comparing robust and accurate machine-learning models with spectral data, texture features and data fusion respectively, to quantitatively identify the tobacco disease; (iv) discriminating TMV-infected from non-diseased tobacco leaves and classifying three levels of disease degree during the infected period even before specific symptoms became visible

Summary

Introduction

With conventional RGB imaging, near infrared (NIR) spectroscopy and multispectral imaging, HSI integrates conventional imaging and spectroscopy to obtain both spatial and spectral information simultaneously from a sample at spatial resolutions varying from the level of single cells up to macroscopic objects[6, 8, 9]. Extreme learning machine model was developed for the classification of healthy, early blight and late blight of detached leaves, which yielded about 97.1–100% classification accuracy when classifying diseased plants based on spectral information. This outmost goal was achieved by meeting the following specific objectives: (i) determining the corresponding effective wavelengths (EWs) which give the highest correlation between the spectral data and different disease stages; (ii) extracting texture features based on grey-level co-occurrence matrix (GLCM) at the selected EWs; (iii) developing and comparing robust and accurate machine-learning models with spectral data, texture features and data fusion respectively, to quantitatively identify the tobacco disease; (iv) discriminating TMV-infected from non-diseased tobacco leaves and classifying three levels of disease degree during the infected period even before specific symptoms became visible

Objectives

Methods

Results

Conclusion