Electronic Nose-Based Technique for Rapid Detection and Recognition of Moldy Apples

Abstract

In this study, the PEN3 electronic nose was used to detect and recognize fresh and moldy apples inoculated with Penicillium expansum and Aspergillus niger, taking Golden Delicious apples as the model subject. Firstly, the apples were divided into two groups: individual apple inoculated only with/without different molds (Group A) and mixed apples of inoculated apples with fresh apples (Group B). Then, the characteristic gas sensors of the PEN3 electronic nose that were most closely correlated with the flavor information of the moldy apples were optimized and determined to simplify the analysis process and improve the accuracy of the results. Four pattern recognition methods, including linear discriminant analysis (LDA), backpropagation neural network (BPNN), support vector machines (SVM), and radial basis function neural network (RBFNN), were applied to analyze the data obtained from the characteristic sensors, aiming at establishing the prediction model of the flavor information and fresh/moldy apples. The results showed that only the gas sensors of W1S, W2S, W5S, W1W, and W2W in the PEN3 electronic nose exhibited a strong signal response to the flavor information, indicating most were closely correlated with the characteristic flavor of apples and thus the data obtained from these characteristic sensors were used for modeling. The results of the four pattern recognition methods showed that BPNN had the best prediction performance for the training and testing sets for both Groups A and B, with prediction accuracies of 96.3% and 90.0% (Group A), 77.7% and 72.0% (Group B), respectively. Therefore, we demonstrate that the PEN3 electronic nose not only effectively detects and recognizes fresh and moldy apples, but also can distinguish apples inoculated with different molds.

A Proprietarily Developed Bionic Olfactory System Used for Rapid Detection of Deteriorated Refrigerated-Stored Apples

“…Researchers usually study fruit hardness, sugar degree, decay, freshness and infectious diseases of bananas [7], oranges [8], cherries [9] and apples [10]- [13] by the electronic nose. The studies prove that the applicability of electronic nose to apple quality detection, while most scholars have focused the studies on apples artificially [14] infected with pathogenic bacteria without giving much attention to naturally deteriorated apples in cold storage. Electronic noses can be classified into two types: the first type is universal electronic noses, which are mainly costly PEN3 [15], [16]; the second type is customized, special-purpose electronic nose.…”

Historical Evolution and Food Control Achievements of Near Infrared Spectroscopy, Electronic Nose, and Electronic Tongue—Critical Overview

“…In one of the recent studies, a portable commercial PEN3 E-nose was used to detect and recognize moldy ( Penicillium expansum , Aspergillus niger ) “Golden Delicious” apples. The data corresponding to the most correlating sensors were processed by LDA, BPNN, SVM, and radial basis function neural network (RBFNN), where the best results were obtained with BPNN and the prediction accuracies varied between 72% and 96.3% [ 199 ]. Additionally, the portable commercial PEN3 was employed to monitor total soluble solid (TSS) and titratable acidity (TA) of litchi fruit exposed to varying degrees of mechanical damage (intact, mild, severe) and different storage conditions (room and cold temperature).…”

Machine Learning for Predicting Mycotoxin Occurrence in Maize

“…DNNs are also very useful for plant disease identification, which is done via convolutional neural networks (CNN), which is a specific DNN-architecture used for image recognition (Boulent et al, 2019). Mycotoxins are mainly detected via high-performance liquid chromatography (HPLC) and mass spectrometry; however, DNNs coupled with rapid analytical tools, such as the electronic nose or infrared attenuated total reflection spectroscopy, have been recently applied and found to improve the assessment reliability (Evans et al, 2000;Jia et al, 2019;Öner et al, 2019;Camardo Leggieri et al, 2020b). Torelli et al (2012), in the first example of ML applied to mycotoxins, performed a 2-year study (2007)(2008) that included seven cropping system variables-FAO class, sowing and harvest dates, crop duration, kernel moisture, ECB treatment, and irrigation-as input for an ANN to classify maize samples based on their contamination with FBs.…”