AI-Based RGB Imaging and 14-Channel Multispectral Food Quality Scanner

Abstract

Food-quality assessment plays an important role in food safety, agricultural processing, storage, and supply-chain management. Conventional inspection methods depend on visual judgement or laboratory-based testing, which may require considerable time, specialised equipment, and manual effort. Recent developments in computer vision, multispectral sensing, and artificial intelligence (AI) provide opportunities for automated, non-destructive food-quality assessment. This paper reviews the application of RGB imaging, multispectral sensing, machine learning, and deep learning for fruit-ripeness estimation, surface-defect detection, grain classification, and food-authenticity assessment. The reviewed literature demonstrates that RGB imaging can extract visible features such as colour, shape, and texture, whereas multispectral sensing provides wavelength-dependent information associated with food characteristics. Published studies have reported promising results in tomato-ripeness classification, apple-defect detection, rice-grain analysis, and honey-adulteration screening. Based on these findings, a compact food-quality scanner integrating an RGB camera, a 14-channel multispectral sensor, controlled illumination, and an embedded processing unit is proposed. The review discusses sensor integration, data preprocessing, feature fusion, classification methods, evaluation metrics, and implementation challenges. It also identifies the need for broader datasets, standardised illumination, independent validation, and resource-efficient AI models to improve the reliability of portable food-inspection systems.

Introduction

The text presents a review and proposed design of an AI-Based Food Quality Scanner Using RGB Imaging and 14-Channel Multispectral Sensing. The research focuses on improving food-quality inspection through computer vision, multispectral measurements, and machine learning. Traditional food inspection relies on manual examination or laboratory testing, which can be subjective, time-consuming, and expensive. The proposed system aims to provide a portable, non-destructive, and relatively low-cost alternative for preliminary food-quality assessment.

The study explains that food quality depends on several characteristics, including colour, shape, size, texture, ripeness, surface defects, and chemical composition. RGB cameras can capture visible features and identify defects such as bruising and discolouration. However, visible appearance alone may not reveal internal damage or subtle compositional differences. Multispectral sensors address this limitation by measuring light intensity at selected wavelengths, providing additional information about food samples.

The literature review examines existing research on RGB-based food inspection, multispectral fruit-ripeness assessment, defect detection, grain-quality analysis, honey adulteration detection, and spectral measurement stability. Previous studies demonstrate promising results, including 93.72% accuracy for multispectral tomato-ripeness classification, 96.9% accuracy for multispectral apple-defect classification, and 100% accuracy for Random Forest-based honey-adulteration detection in one reported dataset. These results demonstrate the potential of optical sensing and machine learning but should not be assumed to represent the performance of the proposed scanner.

The proposed system integrates the following major components:

• RGB Camera: An OV5640 5 MP camera captures visible food characteristics, including colour, shape, texture, and surface defects.

• Multispectral Sensor: The AS7343 14-channel sensor measures wavelength-dependent light intensities to provide additional spectral information.

• Illumination Unit: White LEDs support visible imaging, while suitable near-infrared illumination may support compatible spectral measurements.

• Processing Unit: An ESP32-S3 N16R8 manages sensor communication, data acquisition, and scanning operations, subject to hardware capability testing.

• AI Classification Pipeline: Machine-learning or deep-learning models analyse image and spectral features to classify food categories, estimate ripeness, detect visible defects, and assess quality.

• Display and Storage: A display presents classification results, while a microSD card stores scan records, sensor measurements, and related metadata.

The system operates through a sequence of sample placement, controlled illumination, RGB image capture, multispectral measurement, preprocessing, AI-based classification, and result generation. Both data streams are associated with the same sample so that visible and spectral features can be analysed together. Environmental measurements may also provide contextual information, although they cannot independently establish food freshness.

The research identifies several important limitations in existing approaches, including the inability of RGB imaging to reveal all internal defects, the cost and complexity of hyperspectral systems, differences between multispectral sensors, and limited generalization of models trained on specific food categories. Reliable operation also requires consistent illumination, sensor calibration, controlled sample positioning, and appropriate separation of training and testing datasets.

The principal research contribution is the proposed integration of RGB imaging and 14-channel multispectral sensing into a single scanner. The study aims to compare three approaches: RGB-only classification, multispectral-only classification, and combined-feature classification. This comparison will help determine whether combining visible and spectral information improves food-quality assessment.

Conclusion

This paper reviewed the use of RGB imaging, multispectral sensing, and artificial intelligence for non-destructive food-quality assessment. Existing studies demonstrate the potential of RGB image processing for visible food characteristics, multispectral measurements for fruit-ripeness estimation and food-authenticity screening, and deep-learning models for defect classification. Hyperspectral imaging provides richer spectral information but introduces additional cost and processing complexity. Based on these findings, an AI-based food-quality scanner combining an RGB camera and a 14-channel multispectral sensor has been proposed. The architecture integrates controlled illumination, data acquisition, preprocessing, feature extraction, AI classification, and result storage. The principal research objective is to determine whether combining RGB and spectral information improves performance compared with either sensing method alone. Further experimental work is required to validate sensor integration, establish a representative dataset, compare classification algorithms, and evaluate generalisation under different scanning conditions

References

[1] A. Jatain, S. B. Bajaj, P. Vashisht, and A. Narang, “AI Based Food Quality Recommendation System,” International Journal of Innovative Research in Computer Science & Technology, vol. 11, no. 3, pp. 20–26, May 2023, doi: 10.55524/ijircst.2023.11.3.4. [2] G. M. S. Priyadharshini, A. Garai, and P. Ghosh, “Color-based imaging in post-harvest fruit quality assessment: A comprehensive review,” Frontiers in Sustainable Food Systems, vol. 10, Art. no. 1897830, 2026, doi: 10.3389/fsufs.2026.1897830. [3] K. Sandhya Rani, K. Swetha, K. Amrutha Varshini, G. Harika, D. Trisnawarman, A. A. Supriyanto, and V. C. Mawardi, “Rice Grain Quality Analysis Using Image Processing,” International Journal of Advances in Artificial Intelligence and Machine Learning, vol. 2, no. 2, 2025, doi: 10.58723/ijaaiml.v2i2.455. [4] C. Lauretti et al., “A Low-Cost Multispectral Device for In-Field Fruit Ripening Assessment,” IEEE Sensors Journal, vol. 25, no. 15, pp. 27834–27844, 2025. [5] B. X. Vintimilla, L. E. Chuquimarca, G. S. Yaguana, and E. L. Auria, “Classification of fruit defects using convolutional neural networks and multispectral imaging,” Journal of Agriculture and Food Research, vol. 30, Art. no. 103149, 2026, doi: 10.1016/j.jafr.2026.103149. [6] Y. Liang et al., “Advances in Hyperspectral Imaging Technology for Grain Quality and Safety Detection: A Review,” Foods, vol. 14, no. 17, Art. no. 2977, 2025, doi: 10.3390/foods14172977. [7] V. M. Biradar, S. Hiregoudar, R. Patil, and S. Begum, “Detection of Jaggery Syrup Adulteration in Apis mellifera Honey Using a Multispectral Sensor Coupled with Machine Learning,” European Journal of Nutrition & Food Safety, vol. 18, no. 8, pp. 29–42, 2026, doi: 10.9734/ejnfs/2026/v18i82114. [8] P. Myland, A. Herzog, S. Babilon, W. D. van Driel, and T. Q. Khanh, “LED Degradation Monitoring Using a Multi-Channel Spectral Sensor,” IEEE Access, vol. 12, pp. 42295–42310, 2024, doi: 10.1109/ACCESS.2024.3378101. [9] C. Lauretti et al., “A Low-Cost Multispectral Device for In-Field Fruit Ripening Assessment,” IEEE Sensors Journal, vol. 25, no. 15, pp. 27834–27844, 2025. [10] Y. Fang, H. Bai, L. Sun, J. Hou, Y. Che, Multispectral imaging-based detection of apple bruises using segmentation network and classification model, J. Food Sci. 90 (1) (2025) 70003. [11] P. Pathmanaban, B. Gnanavel, S.S. Anandan, S. Sathiyamurthy, Advancing post-harvest fruit handling through Al-based thermal imaging: applications, challenges, and future trends, Discov. Food 3 (1) (2023) 27. [12] LE. Chuquimarca, B.X. Vintimilla, S.A. Velastin, A review of external quality inspection for fruit grading using CNN models, Artif. Intell. Agric. 14 (2024) 1-20 [13] V. N., A.J. Pinto, Defects detection in fruits and vegetables using image processing and soft computing techniques, 2024, ResearchGate. URL Fruits and Vegetables Using Image Processing and Soft Computing Techniques. [14] Y. Liu, C. Zhang, X. Dong, A survey of real-time surface defect inspection methodsbased on deep learning, Artif. Intell. Rev. 56 (10) (2023) 12131-12170 [15] F. Shahbazi, S. Shahbazi, D. Zare, Losses in agricultural produce: Causes and effects on food security, Food Energy Secur. 14 (3) (2025) 70006.

Copyright

Copyright © 2026 Eshwari Take, Avanish Raut, Vedant Joshi, Shruti Dive, Shravani Akolkar, R. M. Gharat. This is an open access article distributed under the Creative Commons Attribution License, which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited.