Regular articleMeasurement orientation compensation and comparison of transmission spectroscopy for online detection of moldy apple core
Introduction
Apples are widely available in the global market and are popular around the world for their excellent flavor and nutritional value [1], [2]. The ever-increasing consumer awareness of food quality makes it essential for the apple industry to maintain high standards of fruit quality [3], especially because the revenue from such high-quality products is much higher than the average income. However, apples are susceptible to a variety of defects in its pre-harvest, at harvest and during post-harvest operations, such as storage, transport and retail distribution. Therefore, it is important to detect these defects after harvest and before sale.
The defects suffered by apples generally fall into the following categories: physiological disorders, mechanical damage, pathological disorders, and contamination [1]. Pathological disorders and contamination usually cause more serious defects than the other categories, and these defects have seriously endangered food safety, so fruit inspection for these issues is performed using higher and stricter standards. Because of the acidity of the apple tissue, apples are susceptible to fungal infection, which can cause pathological disorders [4], [5], [6], [7], [8]. One of the most common pathological disorders is moldy core, which is characterized by dry, spongy brown lesions in the core area [9], [10], [11]. Preventing such defects is very difficult, and apples with moldy cores cannot be detected until the apples are cut open [12]. Ideally, diseased fruit would be removed prior to storage; otherwise, healthy apples are likely to be infected. Therefore, there is a strong need within the apple industry for a reliable, nondestructive method for detecting and grading such fruit.
Several nondestructive methods for detecting internal defects in apples have been developed. These methods involve X-ray imaging [13], [14], magnetic resonance imaging [15], [16], [17], [18], and thermal imaging [19], but the commercial application of these methods has a big drawback, that is, they are very expensive and time-consuming. Vis/NIR transmission spectroscopy is a fast, cheap, and nondestructive method for the detection of internal apple diseases. This technology was first used to detect internal diseases in apples in the 1980s and is now largely mature [20], [21], [22], [23], [24], [25], [26], [27]. In the research of Upchurch et al. [28], the ability of transmission spectroscopy to detect internal diseases in apples had been demonstrated. After this study, four arrangements of light source, fruit-orientation and detector used to detect the brownheart in apples were investigated by Clark et al. [29]. Guo et al. [30] quantitatively analyzed the internal quality and diseases in apples by using near infrared transmittance spectroscopy and obtained good detection results. These studies focused on the static, rather than the dynamic or online, state of fruit. For commercial applications, this type of research is insufficient, and a technology that can rapidly detect moldy core in an online manner, rather than statically, is urgently required. Shenderey et al. [31] evaluated the ability of Vis-NIR minispectrometers to detect moldy core in apples, on line. This method provides a research basis for online detection in commercial applications. McGlone et al. [32] developed two prototype online transmission spectroscopy systems. The influence of two different spectrometers on the detection results has been investigated in the literature. However, measurement orientation is also a significant factor affecting the quality of the acquired spectra during online detection [33], [34]. Fu et al. [35] used transmission spectra to detect brown heart in pears, and better results were obtained in fruit orientation T2. Han et al. [36] compared three fruit orientations (T1, T2, T3) for detecting the soluble solids content in apples, and obtained better results in fruit orientation T1. Clearly, during the online detection of moldy core in apples, the methods to evaluate the spectral information in different orientations and the methods to eliminate unwanted effects are lacking, and it is necessary for us to study this part of the content.
The general goal of this study was to develop an optimal model to realize the online detection of moldy core in Fuji apples, in terms of fruit orientation. The specific objectives were (1) to investigate the performance of an online system that collects the transmission spectra of apples; (2) to evaluate the stability of spectral information obtained from different orientations using the signal-to-noise ratio (SNR) and area change rate (ACR); and (3) to establish and compare the moldy core detection models generated by the local orientation and the global orientation.
Section snippets
Samples
To better describe our samples, we refer to the apples with a moldy core as diseased apples and the apples with a normal core as healthy apples (Fig. 1). The research object were red Fuji apples. Using portable equipment developed in our laboratory [37], 132 suspected diseased apples and 123 suspected healthy apples were selected for subsequent experimental analysis. All 255 Fuji apples were collected in late September 2019 from the same refrigerated storage (37°25′5.63″N, 120°53′5.84″E, 82 m
Overview of spectra and sample data distribution
The raw transmission spectrum obtained by the system in the T1 orientation was shown in Fig. 5. Observations of the spectrum revealed two distinct absorption peaks at 640 nm and 714 nm. The spectra obtained in orientations T2 and T3 had similar characteristics, and their spectral variation trends were also the same. However, such spectra are useless for detecting moldy core because they contain substantial dark noise information and baseline shifts. The method based on principal component
Conclusions
The influence of variations in spectral measurement orientation on the online detection of moldy core in apples using transmission spectroscopy was studied using a self-designed online spectral measurement system. Through SNR and ANR analysis, we concluded that the spectral information obtained by the system in the T2 orientation was the most stable. After obtaining spectral information in orientations T1, T2, and T3, local and global SVM models were established after SGS and normalization
Declaration of Competing Interest
The authors declare that they have no known competing financial interests or personal relationships that could have appeared to influence the work reported in this paper.
Acknowledgements
This study was financially supported by the Key Research and Development Program of Shaanxi (Item No. 2017ZDXM-NY-017), the National Natural Science Foundation of China (Grant No. 31701664), the Horizontal Scientific Research Project (Item No. ZY-XG-CG-2019-07-06), and the Science and Technology Coordination Innovation Project of Shaanxi (Item No. 2016KTCQ02-14) www.liwenbianji.cn/ac.
References (50)
- et al.
Grading of apples based on firmness and soluble solids content using Vis/SWNIR spectroscopy and spectral scattering techniques
J. Food Eng.
(2014) - et al.
The case for fruit quality: an interpretive review of consumer attitudes, and preferences for apples
Postharvest Biol. Technol.
(2003) - et al.
Effective modification through transmission Vis/NIR spectra affected by fruit size to improve the prediction of moldy apple core
Infrared Phys. Technol.
(2019) - et al.
Characterisation of ‘Braeburn’ browning disorder by means of x-ray micro-CT, Postharvest
Biol. Technol.
(2013) - et al.
X-ray CT for quantitative food microstructure engineering: the apple case
Nucl. Instrum. Meth. B.
(2014) - et al.
Rapid sensing of internal browning in whole apples using a low-cost, low-field proton magnetic resonance sensor
Postharvest Biol. Technol.
(2005) - et al.
Effects of internal browning and watercore on low field (5.4 MHz) proton magnetic resonance measurements of T2 values of whole apples
Postharvest Biol. Technol.
(2008) - et al.
Application of MRI for tissue characterisation of ‘Braeburn’ apple
Postharvest Biol. Technol.
(2013) - et al.
Detection and monitoring of internal browning development in ‘Fuji’ apples using MRI
Postharvest Biol. Technol.
(2001) - et al.
Detection of watercore in ‘Gloster’ apples using thermography
Postharvest Biol. Technol.
(2008)