A new study demonstrates that paper bagging significantly enhances Fuji apple quality and appearance. Hyperspectral imaging combined with machine learning offers a powerful, non-destructive method for evaluating fruit grown under different cultivation conditions.
A recent study recently investigated the type of bagging used during apple cultivation and how that affects the quality of apples using hyperspectral imaging combined with machine learning (ML). This study, which was published in the journal Spectrochimica Acta Part A: Molecular and Biomolecular Spectroscopy, showcased the application of visible–near infrared hyperspectral imaging (HSI), spectral data algorithms, and ML to non-destructively assess apple quality under different cultivation conditions (1). The findings here show how a fruit’s internal and external qualities are impacted by bagging conditions.
Raw Red Fuji Apples | Image Credit: © Brent Hofacker - stock.adobe.com
Previous studies have examined bagging methods and how they impact apple quality (2). In the study, the research team examined 307 Fuji apples in three different types of bags. The conditions tested in this study were unbagged, mesh-bagged, and paper-bagged (1). These bagging types were tested to explore how they affect key quality indicators such as fruit color, size, firmness, soluble solids content (SSC), and aroma profile. By integrating traditional chemical analysis methods with HSI-based modeling, the research team demonstrated a new approach to evaluate fruit quality (1).
Out of the three bagging conditions that were tested, the researchers found that paper-bagged apples possessed the best visual appeal and firmness (1). They demonstrated the largest average diameter and a noticeably brighter red hue compared to their mesh-bagged and unbagged counterparts. Despite having a lower SSC, which is a measure often associated with sweetness, paper-bagged apples had a richer and more complex aroma profile than mesh-bagged fruits (1). Meanwhile, unbagged apples performed the worst, lacking in firmness and color quality.
This comparison revealed statistically significant differences across all quality indicators, highlighting the crucial role of bagging strategies in fruit development.
To interpret the rich spectral data from the HSI system, the team evaluated several preprocessing techniques. The preprocessing techniques examined included moving average smoothing (MAS), Savitzky–Golay (SG), standard normal variate (SNV), multiplicative scatter correction (MSC), and wavelet transform (WT) (1). Out of all these preprocessing techniques, MAS stood out as the most effective method for enhancing spectral clarity and consistency.
In addition, five advanced wavelength selection algorithms, which included successive projections algorithm (SPA), genetic algorithm-partial least squares (GAPLS), competitive adaptive reweighted sampling (CARS), uninformative variable elimination (UVE), and UVE-SPA, were tested to determine which feature wavelengths best predicted quality indicators (1). Out of all these five methods, the CARS algorithm performed the best, demonstrating predictive power for key traits such as color values (L*, a*, b*) and SSC (1).
Using CARS-selected wavelengths, the researchers then built regression models with various ML algorithms. These ML algorithms included partial least squares regression (PLSR), multiple linear regression (MLR), support vector regression (SVR), extreme learning machine (ELM), and back propagation neural network (BPNN) (1). Among these, the CARS-MLR model delivered the highest prediction accuracy, with R²p values above 0.75 for all indicators, making it the most robust and reliable model in the study (1).
The implications of this study are that it provides a theoretical and practical framework for assessing quality in other fruits subjected to varied environmental or agricultural environments. It also offers a scalable solution for commercial quality control, potentially transforming how fruits are evaluated in packing houses and supply chains (1).
The authors acknowledge some limitations. The study focused solely on the Fuji apple variety and collected data from a single growth stage. For their model to be more applicable to other fruits, it needs to be tested with other fruits. However, by demonstrating how bagging choices influence not just aesthetic appeal but also flavor and texture, the study paves the way for more informed decisions by growers and distributors alike (1).
Hyperspectral Imaging for Walnut Quality Assessment and Shelf-Life Classification
June 12th 2025Researchers from Hebei University and Hebei University of Engineering have developed a hyperspectral imaging method combined with data fusion and machine learning to accurately and non-destructively assess walnut quality and classify storage periods.
Machine Learning and NMR Unite to Authenticate Wine with Near-Perfect Accuracy
June 11th 2025In a recent study published in the journal Beverages, a team of researchers from the National Institute for Research and Development of Isotopic and Molecular Technologies and Babeș-Bolyai University explored a new way to improve wine authentication
MIR Spectroscopy Validates Origin of Premium Brazilian Cachaças
June 11th 2025A recent study published in the journal Food Chemistry explored Brazil’s cachaça industry, focusing on a new analytical method that can confirm the geographic origin of cachaças from the Brejo Paraibano region in Brazil.
Harnessing Near-Infrared Spectroscopy and Machine Learning to Detect Microplastics in Chicken Feed
June 5th 2025Researchers from Tianjin Agricultural University, Nankai University, and Zhejiang A&F University have developed a highly accurate method using near-infrared spectroscopy and machine learning to rapidly detect and classify microplastics in chicken feed.
