Measuring the Level of Aflatoxin Infection in Pistachio Nuts by Applying Machine Learning Techniques to Hyperspectral Images

Aug 3, 2026

Upload by Asher

This paper investigates using machine learning on hyperspectral images of pistachios to detect and classify aflatoxin contamination levels. Aflatoxins are toxic mycotoxins produced by moulds and they threaten consumer health, while existing testing approaches are invasive and generate food waste. The study evaluates non-invasive hyperspectral imaging combined with models such as dimensionality reduction with K-Means, residual networks (ResNets), variational autoencoders (VAEs), and deep convolutional GANs (DCGANs). Experiments on 300 images from three contamination ranges identify key wavelengths and report accuracies up to 96.67%, supporting improved pistachio quality control and food safety.

LanguageEnglish

CategoryResearch & Report

FormatPDF

Pages38

Reading Time1 h 36 min

AccessFree to download

Introduction

Background and health risk

Regulations for aflatoxin levels

Motivation: reducing waste with non-invasive testing

Methods

Hyperspectral imaging pipeline

Data and contamination levels

Model set: K-Means, ResNets, VAEs, DCGANs

Feature/wavelength identification

Results and Evaluation

Dimensionality reduction performance

ResNet performance by key wavelength

Impact of dataset size on generative models

Conclusion and Future Work

Potential for quality control

Expanding datasets and refining models

Citation: Williams, L and Shukla, S and Sheikh-Akbari, A and Mahroughi, S and Mporas, I (2025) Measuring the Level of Aflatoxin Infection in Pistachio Nuts by Applying Machine Learning Techniques to Hyperspectral Images. Sensors, 25 (5). pp. 1-37. ISSN 1424-2818 DOI: [https://doi.org/10.3390/s25051548](https://doi.org/10.3390/s25051548) Link to Leeds Beckett Repository record: [https://eprints.leedsbeckett.ac.uk/id/eprint/11854/](https://eprints.leedsbeckett.ac.uk/id/eprint/11854/) Document Version: Article (Published Version) Creative Commons: Attribution 4.0 © 2025 by the authors The aim of the Leeds Beckett Repository is to provide open access to our research, as required by funder policies and permitted by publishers and copyright law. The Leeds Beckett repository holds a wide range of publications, each of which has been checked for copyright and the relevant embargo period has been applied by the Research Services team. We operate on a standard take-down policy. If you are the author or publisher of an output and you would like it removed from the repository, please contact us and we will investigate on a case-by-case basis. Each thesis in the repository has been cleared where necessary by the author for third party copyright. If you would like a thesis to be removed from the repository or believe there is an issue with copyright, please contact us on [openaccess@leedsbeckett.ac.uk](openaccess@leedsbeckett.ac.uk) and we will investigate on a case-by-case basis. Article Measuring the Level of Aflatoxin Infection in Pistachio Nuts by Applying Machine Learning Techniques to Hyperspectral Images Lizzie Williams 1, Pancham Shukla 1, Akbar Sheikh-Akbari 2, *, Sina Mahroughi 2 and Iosif Mporas 3 Academic Editor: Subrata Chakraborty Received: 24 November 2024 Revised: 19 February 2025 Accepted: 27 February 2025 Published: 2 March 2025 Citation: Williams, L.; Shukla, P.; Sheikh-Akbari, A.; Mahroughi, S; Mporas, I. Measuring the Level of Aflatoxin Infection in Pistachio Nuts by Applying Machine Learning Techniques to Hyperspectral Images. Sensors 2025, 25, 1548. [https://](https://)[ ](https://)[doi.org/10.3390/s25051548](doi.org/10.3390/s25051548) [Copyright:](Copyright:) © 2025 by the authors. Licensee MDPI, Basel, Switzerland. This article is an open access article distributed under the terms and conditions of the Creative Commons Attribution (CC BY) license ([https://creativecommons.org/](https://creativecommons.org/)[ ](https://creativecommons.org/)[licenses/by/4.0/](licenses/by/4.0/)) . 1 Department of Computing, Imperial College London, London SW7 2AZ, UK; [lizzies.williams02@gmail.com](lizzies.williams02@gmail.com) (L.W.); [panchamkumar.shukla@imperial.ac.uk](panchamkumar.shukla@imperial.ac.uk) (P.S.) 2 School of Built Environment, Engineering and Computing, Leeds Beckett University, Leeds LS6 3QS, UK; [s.mahroughi@leedsbeckett.ac.uk](s.mahroughi@leedsbeckett.ac.uk) 3 Department of Engineering and Technology, School of Physics, Engineering & Computer Science, University of Hertfordshire, Hatfield AL10 9AB, UK; i.mporas@herts.ac.uk * Correspondence: [a.sheikh-akbari@leedsbeckett.ac.uk](a.sheikh-akbari@leedsbeckett.ac.uk); Tel.: +44-(0)1138121767 Abstract: This paper investigates the use of machine learning techniques on hyperspectral images of pistachios to detect and classify different levels of aflatoxin contamination. Aflatoxins are toxic compounds produced by moulds, posing health risks to consumers. Current detection methods are invasive and contribute to food waste. This paper explores the feasibility of a non-invasive method using hyperspectral imaging and machine learning to classify aflatoxin levels accurately, potentially reducing waste and enhancing food safety. Hyperspectral imaging with machine learning has shown promise in food quality control. The paper evaluates models including Dimensionality Reduction with K-Means Clustering, Residual Networks (ResNets),

FAQs

What problem does the paper address regarding pistachios?

The paper targets detecting and classifying different levels of aflatoxin contamination in pistachios, which are harmful toxins produced by moulds.

Why are hyperspectral imaging and machine learning used together?

The approach aims to provide a non-invasive alternative to current invasive detection methods, using hyperspectral imaging features and machine learning models to classify contamination levels accurately.

Which models and key performance results are reported?

The study evaluates K-Means-based dimensionality reduction, ResNets, VAEs, and DCGANs. Dimensionality reduction with K-Means reaches 84.38% accuracy, while a ResNet model using the 866.21 nm wavelength achieves 96.67%.

Aug 3, 2026

Upload by Asher