Classification of Pomelo Maturity Using Low-Cost Peel Feature Analysis and Machine Learning Algorithms
Article Sidebar
Main Article Content
Abstract
The lack of accurate pre-harvest maturity assessment tools is a critical challenge for pomelo production. Since pomelo is a non-climacteric fruit, its internal quality becomes fixed immediately upon harvest. Consequently, inaccuracies in harvest timing decisions directly result in economic losses. This study presents a cost-effective machine learning framework for maturity classification (early-mature vs. mature) of Khao Nampueng pomelos. A balanced dataset of 1,008 images from 70 pomelos was obtained using a Raspberry Pi camera at two critical times: 180 and 210 Days After Fruit Set (DAFS). To capture the non-linear relationships between peel appearance and maturity, color and texture features were analyzed using the Permutation Feature Importance method integrated with a Support Vector Machine (SVM) utilizing a Radial Basis Function (RBF) kernel. The proposed method yielded an optimized subset of biologically relevant features. The SVM-RBF model trained on this subset achieved robust classification performance, with a test accuracy of 95.6%. Interpretability analyses confirmed that the model’s decision boundary aligned with physiological and biological changes, identifying variation in the chromatic component (b* channel) and micro-texture (LBP) as the most significant predictors. This study demonstrates that cost-effective peel feature analysis driven by machine learning algorithms provides a high-precision tool for optimizing pomelo harvest timing.
Article Details
This work is licensed under a Creative Commons Attribution-NonCommercial-NoDerivatives 4.0 International License.
The articles published are the opinion of the author only. The author is responsible for any legal consequences. That may arise from that article.
References
J. Yin, X. Hu, Y. Hou, S. Liu, S. Jia, C. Gan, Y. Ou, and X. Zhang, “Comparative analysis of chemical compositions and antioxidant activities of different pomelo varieties from China,” Food Chemistry Advances, vol. 2, 2023, Art. no. 100180, doi: 10.1016/j.focha. 2022.100180.
T. Sarakum and S. Sukpancharoen, “Nondestructive sweetness classification of Khao Tang Kwa pomelos using machine learning with acoustic and image processing,” Journal of Food Composition and Analysis, vol. 142, 2025, Art. no. 107385, doi: 10.1016/j.jfca.2025.107385.
K. Chawgien and S. Kiattisin, “Machine learning techniques for classifying the sweetness of watermelon using acoustic signal and image processing,” Computers and Electronics in Agriculture, vol. 181, 2021, Art. no. 105938, doi: 10.1016/j.compag.2020.105938.
B. Lu, P. D. Dao, J. Liu, Y. He, and J. Shang, “Recent advances of hyperspectral imaging technology and applications in agriculture,” Remote Sensing, vol. 12, no. 16, pp. 2659, 2020, doi: 10.3390/rs12162659.
A. Koirala, K. B. Walsh, Z. Wang, and C. Mccarthy, “Deep learning – method overview and review of use for fruit detection and yield estimation,” Computers and Electronics in Agriculture,vol. 162, pp. 219–234, 2019, doi: 10.1016/ j.compag.2019.04.017.
X. Guan, H. Wan, Z. He, Z. Liu, R. Jiang, Y. Ou, Y. Chen, H. Gu, and Z. Zhou, “Pomelo-Net: A lightweight semantic segmentation model for key elements segmentation in honey pomelo orchard for automated navigation,” Computers and Electronics in Agriculture, vol. 229, 2025, Art. no. 109760, doi: 10.1016/ j.compag.2024.109760.
D. Malowany and H. Guterman, “Biologically inspired visual system architecture for object recognition in autonomous systems,” Algorithms, vol. 13, no. 7, pp. 167, 2020, doi: 10.3390/ a13070167.
H. Chen, H. Qiao, L. Xu, Q. Feng, and K. Cai, “A fuzzy optimization strategy for the implementation of RBF LSSVR model in Vis-NIR analysis of pomelo maturity,” IEEE Transactions on Industrial Informatics, vol. 15, no. 11, pp. 5971–5979, 2019, doi: 10.1109/ TII.2019.2933582.
T. Jiang, J. Ding, Y. Du, S. Yuan, H. Yu, and W. Yao, “Deep learning-driven Vis/NIR spectroscopic devices for fruit quality assessment: A comprehensive review,” Trends in Food Science and Technology, vol. 164, 2025, Art. no. 105262, doi: 10.1016/j.tifs.2025.105262.
N. Upadhyay and A. Bhargava, “Artificial intelligence in agriculture : Applications, approaches , and adversities across pre-harvesting, harvesting, and post-harvesting phases,” Iran Journal of Computer Science, vol. 8, no. 3, pp. 749–772, 2025, doi: 10.1007/s42044-025- 00264-6.
H. Wang, M. Mei, and J. Li, “Research progress on non-destructive detection of internal quality of fruits with large size and thick peel : A review,” Agriculture, vol. 13, no. 9, pp. 1838, 2023, doi: 10.3390/agriculture13091838.
S. S. Harakannanavar, J. M. Rudagi, V. I. Puranikmath, A. Siddiqua, and R. Pramodhini, “Plant leaf disease identification method using computer vision and machine learning algorithms,” International Conference on Integrated Intelligence and Communication Systems, ICIICS 2023, vol. 3, pp. 305–310, 2023, doi: 10.1109/ICIICS59993.2023.10421345.
G. Huang, Z. Xu, L. Liu, J. Chen, Z. He, and F. Liu, “Evaluating and deploying large vision-language models for fruit quality assessment in smart agriculture systems,” Computers and Electronics in Agriculture, vol. 238, 2025, Art. no. 110806, doi: 10.1016/j.compag.2025.110806.
E. Sheidaee and P. Bazyar, “Detection of white fig ripeness stages using deep learning models,” Iran Journal of Computer Science, vol. 8, pp. 2935–2498, 2025, doi: 10.1007/s42044- 025-00307-y.
S. A. E. Kabel, W. El-shafai, F. E. A. El-samie, and B. Mohamed, “Enhanced plant identification and disease diagnosis through SqueezeNet and SVM for smart agriculture applications,” Iran Journal of Computer Science, vol. 8, pp. 1353– 1369, 2025, doi: 10.1007/s42044-025-00237-9.
M. Soltani and F. Hamed, “Defect detection in fruit and vegetables by using machine vision systems and image processing,” Food Engineering Reviews, vol. 14, no. 3, pp. 353–379, 2022, doi:10.1007/s12393-022-09307-1.
A. K. Gupta, U. Pathak, T. Tongbram, M. Medhi, A. Terdwongworakul, L. S. Magwaza, A. Mditshwa, T. Chen, and P. Mishra, “Emerging approaches to determine maturity of citrus fruit,” Critical Reviews in Food Science and Nutrition, vol. 62, no. 19, pp. 5245–5266, 2022, doi: 10.1080/10408398.2021.1883547.
C. Wang, W. Suk, L. Xiangjun, Z. Daeun, and C. Hao, “Detection and counting of immature green citrus fruit based on the Local Binary Patterns (LBP) feature using illumination ‑ normalized images,” Precision Agriculture, vol. 19, no. 6, pp. 1062–1083, 2018, doi: 10.1007/s11119- 018-9574-5.
J. Neyman, “On the two different aspects of the representative method: The method of stratified sampling and the method of purposive selection,” Journal of the Royal Statistical Society, vol. 97, no. 4, pp. 558–625, 1934, doi: 10.2307/2342192.
E. Junda, C. Málaga-Chuquitaype, and K. Chawgien, “Interpretable machine learning models for the estimation of seismic drifts in CLT buildings,” Journal of Building Engineering, vol. 70, 2023, Art. no. 106365, doi: 10.1016/j. jobe.2023.106365.
E. Junda and C. Málaga-Chuquitaype, “Seismic acceleration demands in tall CLT buildings, predictive models and intensity measures,” Engineering Structures, vol. 298, 2024, Art. no. 117024, doi: 10.1016/j.engstruct. 2023.117024.
S. Masís, Interpretable Machine Learning with Python. 2022, doi:10.1142/12774.
