Nondestructive Determination of Epicarp Hardness of Passion Fruit Using Near-Infrared Spectroscopy during Storage
Abstract
:1. Introduction
2. Materials and Methods
2.1. Experimental Sample
2.2. Near-Infrared Spectra Acquisition
2.3. Texture Measurement
2.4. Multivariate Data Analysis
2.4.1. NIRS Sample Set Partitioning
2.4.2. NIRS Data Preprocessing
2.4.3. Effective Wavelength Selection
2.4.4. Random Forest Model of Grid Search Algorithm Optimization
- (1)
- Grid search is employed to explore a wide range of combination values of the parameters “ntree” and “mtry”.
- (2)
- K-fold cross-validation, a fundamental element in the RF training process, is used to calculate the average error of the k results. This average error is referred to as the cross-validation error of the RF model.
- (3)
- An iterative process is initiated across the “ntree” and “mtry” parameters, systematically applying varying combinations of parameter values for RF training and simultaneously calculating the cross-validation error for each iteration.
- (4)
- The RF model associated with the smallest cross-validation error is selected as the optimized model.
2.4.5. Support Vector Regression Model of Genetic Algorithm Optimization
- (1)
- An initial population is constructed, which encompasses a diverse array of parameter combinations for “c” and “γ”.
- (2)
- The pre-configured settings are applied to each corresponding set of “c” and “γ” for training the SVR model. The mean squared error (MSE) is formulated as the fitness measure through the calculation:
- (3)
- Samples of high fitness are selected, advantageous characteristics from multiple samples are integrated using crossover, and unpredictability is injected by performing mutation operations, preventing the model from falling into local optima. The population size is kept consistent by replacing lower-fitness samples with new ones.
- (4)
- The optimization process—including the preservation of a consistent population size, the crossover and mutation operations, as well as selection mechanisms—is repeated until the fitness measure stabilizes and no longer decreases.
- (5)
- This iterative process results in the emergence of “c” and “γ” values associated with the minimal MSE.
- (6)
- The final optimized SVR model is then obtained by using the “c” and “γ” values determined in Step 5.
2.4.6. Model Evaluation Indicators
3. Results
3.1. Analysis of Patterns of Variation in Textural Parameters
3.2. Result of Set Partitioning
3.3. Results of Spectra Preprocessing
3.4. Discussion of Effective Wavelength Selection
3.5. Discussion of Final Models
4. Conclusions
Author Contributions
Funding
Institutional Review Board Statement
Informed Consent Statement
Data Availability Statement
Conflicts of Interest
References
- De Jesus, O.N.; Lima, L.K.S.; Soares, T.L.; da Silva, L.N.; dos Santos, I.S.; Sampaio, S.R.; de Oliveira, E.J. Phenotypic diversity and alternative methods for characterization and prediction of pulp yield in passion fruit (Passiflora spp.) germplasm. Sci. Hortic. 2022, 292, 110573. [Google Scholar] [CrossRef]
- Zhong, Z.; Zhou, L.; Yu, K.; Jiang, F.; Xu, J.; Zou, L.; Du, L.; Liu, W. Effects of microporous packaging combined with chitosan coating on the quality and physiological metabolism of passion fruit after harvest. Food Bioprocess Technol. 2022, 15, 1836–1850. [Google Scholar] [CrossRef]
- Nxumalo, K.A.; Fawole, O.A. Effects of chitosan coatings fused with medicinal plant extracts on postharvest quality and storage stability of purple passion fruit (passiflora edulis var.ester). Food Qual. Safety 2022, 6, 19. [Google Scholar] [CrossRef]
- Owino, W.; Shibairo, S.; Ambuko, J.; Yumbya, P. Effect of modified atmosphere packaging (map) on the shelf life and postharvest quality of purple passion fruit (passiflora edulis sims). J. Post-Harvest. Technol. 2014, 2, 25–36. [Google Scholar]
- Yanping, Y.; Cao, C.; Yi, Y.; Jiang, J.; Bo, B.; Li, L.; Feng, F. Biomechanical modeling of surface wrinkling of soft tissues with growth-dependent mechanical properties. Acta Mech. Solida Sin. 2012, 25, 483–492. [Google Scholar]
- Wei, X.; He, J.; Zheng, S.; Ye, D. Modeling for SSC and firmness detection of persimmon based on NIR hyperspectral imaging by sample partitioning and variables selection. Infrared Phys. Technol. 2019, 105, 103099. [Google Scholar] [CrossRef]
- Oh, T.G.; Jo, J.A.; Lee, S.J. Evaluation of time–temperature integrator for indicating the ripeness of kiwifruit in plastic container at home. J. Food Sci. 2021, 86, 2872–2885. [Google Scholar] [CrossRef] [PubMed]
- Salvador, A.; Arnal, L.; Besada, C.; Larrea, V.; Quiles, A.; Pérez-Munuera, I. Physiological and structural changes during ripening and deastringency treatment of persimmon fruit cv. ‘Rojo Brillante.’ Postharvest Biol. Technol. 2007, 46, 181–188. [Google Scholar] [CrossRef]
- Maniwara, P.; Nakano, K.; Boonyakiat, D.; Ohashi, S.; Hiroi, M.; Tohyama, T. The use of visible and near infrared spectroscopy for evaluating passion fruit postharvest quality. J. Food Eng. 2014, 143, 33–43. [Google Scholar] [CrossRef]
- Liu, H. Changes in Cell Wall Components and Degrading Enzyme Activity and Their Relation to Hardness in Fruits of Peach (Prunus persica L.) before and after Ripening. Plant Physiol. Commun. 2007, 43, 837. [Google Scholar]
- Júnior, A.; Tan, F.; Mo, X.; Ruan, S.; Yan, T.; Peng, X.; Liu, T. Combining Vis-NIR and NIR Spectral Imaging Techniques with Data Fusion for Rapid and Nondestructive Multi-Quality Detection of Cherry Tomatoes. Foods 2023, 12, 3621. [Google Scholar]
- Sun, J.; Künnemeyer, R.; McGlone, A.; Rowe, P. Multispectral scattering imaging and nir interactance for apple firmness predictions. Postharvest Biol. Technol. 2016, 119, 58–68. [Google Scholar] [CrossRef]
- Ping, F.; Yang, J.; Zhou, X.; Su, Y.; Ju, Y.; Fang, Y.; Liu, W. Quality Assessment and Ripeness Prediction of Table Grapes Using Visible–Near-Infrared Spectroscopy. Foods 2023, 12, 2364. [Google Scholar] [CrossRef]
- Tatelbaum. A Standard Method to Characterize Texture Attributes of Fresh and Processed Foods. Texture Analyzers for Food Quality Evaluation. Texas A&M University. Available online: https://voyagermis.library.tamu.edu/handle/1969.1/157615 (accessed on 21 September 2023).
- Sirisomboon, P.; Pornchaloempong, P. Instrumental textural properties of mango (cv Nam Doc Mai) at commercial harvesting time. Int. J. Food Prop. 2011, 14, 441–449. [Google Scholar] [CrossRef]
- Sirisomboon, P.; Tanaka, M.; Kojima, T. Evaluation of tomato textural mechanical properties. J. Food Eng. 2012, 111, 618–624. [Google Scholar] [CrossRef]
- Sirisomboon, P.; Tanaka, M.; Kojima, T.; Williams, P. Nondestructive estimation of maturity and textural properties on tomato ‘Momotaro’ by infrared spectroscopy. J. Food Eng. 2012, 112, 218–226. [Google Scholar] [CrossRef]
- Fawole, O.A.; Opara, U.L. Harvest discrimination of pomegranate fruit: Postharvest quality changes and relationships between instrumental and sensory attributes during shelf life. J. Food Sci. 2013, 78, S1264–S1272. [Google Scholar] [CrossRef] [PubMed]
- Zhang, W.; Kasun, L.C.; Wang, Q.J.; Zheng, Y.; Lin, Z. A Review of Machine Learning for Near-Infrared Spectroscopy. Sensors 2022, 22, 9764. [Google Scholar] [CrossRef] [PubMed]
- Maniwara, P.; Boonyakiat, D.; Poonlarp, P.; Natwichai, J.; Nakano, K. Changes of postharvest quality in passion fruit (Passiflora edulis Sims) under modified atmosphere packaging conditions. Int. Food Res. J. 2015, 22, 1596–1606. [Google Scholar]
- Xue, J.; Zhang, S.; Sun, H.; Zhou, J. Study of Malus Asiatica Nakai’s firmness during different shelf lives based on visible/near-infrared spectroscopy. Math. Comput. Model. 2013, 58, 1829–1836. [Google Scholar] [CrossRef]
- De Oliveira, C.F.; Gurak, P.D.; Cladera-Olivera, F.; Marczak, L.D.F.; Karwe, M. Combined Effect of High-Pressure and Conventional Heating on Pectin Extraction from Passion Fruit Peel. Food Bioprocess Technol. 2016, 9, 1021–1030. [Google Scholar] [CrossRef]
- Chen, Q.; Ke, G.; Hu, Y.; Fei, P.; Wu, J. Preparation and application of holocellulose, cellulose nanofibers, and silver-loaded cellulose nanofibers from passion fruit peel. Cellulose 2023, 30, 1437–1448. [Google Scholar] [CrossRef]
- de Medeiros, D.T.; Ramalho, F.M.G.; Batista, F.G.; Mascarenhas, A.R.P.; Chaix, G.; Hein, P.R.G. Water desorption monitoring of cellulose pulps by NIR spectroscopy. Ind. Crops Prod. 2023, 192, 115989. [Google Scholar] [CrossRef]
Variable | Minimum/gN | Maximum/gN | Mean/gN | Standard Deviation/gN | Number of Valid Cases/gN |
---|---|---|---|---|---|
F1 | 0.881 | 3.721 | 1.583 | 0.574 | 120 |
F2 | 2.524 | 4.201 | 3.464 | 0.356 | 120 |
Variable | Calibration Set | Prediction Set | ||||||
---|---|---|---|---|---|---|---|---|
Minimum/gN | Maximum/gN | Mean/gN | Quantity | Minimum/gN | Maximum/gN | Mean/gN | Quantity | |
F1 | 0.881 | 3.721 | 1.548 | 90 | 1.064 | 3.544 | 1.646 | 30 |
F2 | 2.524 | 4.201 | 3.432 | 90 | 3.050 | 4.058 | 3.557 | 30 |
Variable | Model | Preprocessing | Calibration Set | Validation Set | RPD | ||
---|---|---|---|---|---|---|---|
R2C | RMSEC/gN | R2P | RMSEP/gN | ||||
F1 | Grids-RF | RAW | 0.869 | 0.208 | 0.874 | 0.211 | 2.706 |
(5 folds) | MC | 0.883 | 0.198 | 0.898 | 0.19 | 2.867 | |
MA | 0.887 | 0.203 | 0.878 | 0.216 | 2.637 | ||
GA-SVR | RAW | 0.831 | 0.085 | 0.841 | 0.084 | 2.141 | |
MC | 0.830 | 0.085 | 0.845 | 0.085 | 2.122 | ||
MA | 0.887 | 0.069 | 0.821 | 0.091 | 1.84 | ||
F2 | Grids-RF | RAW | 0.899 | 0.142 | 0.903 | 0.098 | 2.355 |
(5 folds) | MC | 0.797 | 0.167 | 0.743 | 0.142 | 1.576 | |
MA | 0.912 | 0.127 | 0.908 | 0.101 | 2.379 | ||
GA-SVR | RAW | 0.817 | 0.1 | 0.834 | 0.079 | 2.003 | |
MC | 0.747 | 0.118 | 0.777 | 0.079 | 2 | ||
MA | 0.815 | 0.1 | 0.865 | 0.068 | 2.079 |
Variable | Model | Method | Calibration Set | Validation Set | RPD | ||
---|---|---|---|---|---|---|---|
R2C | RMSEC/gN | R2P | RMSEP/gN | ||||
F1 | Grids-RF (5folds) | MC-CARS | 0.882 | 0.207 | 0.925 | 0.166 | 3.160 |
MC-UVE | 0.888 | 0.202 | 0.903 | 0.184 | 2.781 | ||
MC-SPA | 0.862 | 0.218 | 0.858 | 0.222 | 2.420 | ||
GA-SVR | MC-CARS | 0.897 | 0.143 | 0.901 | 0.075 | 2.773 | |
MC-UVE | 0.890 | 0.144 | 0.818 | 0.122 | 2.012 | ||
MC--SPA | 0.909 | 0.133 | 0.744 | 0.158 | 1.411 | ||
F2 | Grids-RF (5folds) | MA-CARS | 0.882 | 0.147 | 0.868 | 0.110 | 2.209 |
MA-UVE | 0.895 | 0.143 | 0.877 | 0.109 | 2.290 | ||
MA-SPA | 0.833 | 0.168 | 0.746 | 0.155 | 1.208 | ||
GA-SVR | MA-CARS | 0.764 | 0.111 | 0.793 | 0.078 | 1.784 | |
MA-UVE | 0.728 | 0.119 | 0.748 | 0.087 | 1.576 | ||
MA-SPA | 0.738 | 0.105 | 0.717 | 0.102 | 1.479 |
Dependent Variable | Model | Calibration Set | Validation Set | RPD | ||
---|---|---|---|---|---|---|
R2C | RMSEC/gN | R2P | RMSEP/gN | |||
F1 | MC-CARS-Grids-RF (5folds) | 0.882 | 0.207 | 0.925 | 0.166 | 3.160 |
F2 | MA-Grids-RF (5folds) | 0.912 | 0.127 | 0.908 | 0.101 | 2.379 |
Disclaimer/Publisher’s Note: The statements, opinions and data contained in all publications are solely those of the individual author(s) and contributor(s) and not of MDPI and/or the editor(s). MDPI and/or the editor(s) disclaim responsibility for any injury to people or property resulting from any ideas, methods, instructions or products referred to in the content. |
© 2024 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/licenses/by/4.0/).
Share and Cite
Wang, J.; Fu, D.; Hu, Z.; Chen, Y.; Li, B. Nondestructive Determination of Epicarp Hardness of Passion Fruit Using Near-Infrared Spectroscopy during Storage. Foods 2024, 13, 783. https://doi.org/10.3390/foods13050783
Wang J, Fu D, Hu Z, Chen Y, Li B. Nondestructive Determination of Epicarp Hardness of Passion Fruit Using Near-Infrared Spectroscopy during Storage. Foods. 2024; 13(5):783. https://doi.org/10.3390/foods13050783
Chicago/Turabian StyleWang, Junyi, Dandan Fu, Zhigang Hu, Yan Chen, and Bin Li. 2024. "Nondestructive Determination of Epicarp Hardness of Passion Fruit Using Near-Infrared Spectroscopy during Storage" Foods 13, no. 5: 783. https://doi.org/10.3390/foods13050783