International Journal of
Advanced Multidisciplinary Application

Volume 3, Issue 7

• Author(s)Meera Subramaniam
• AffiliationDepartment of Agricultural and Food Engineering, Indian Institute of Technology Kharagpur, West Bengal, India
• Page No.7-13
• Volume, Issue & YearVolume 3, Issue 7, July 2026
• Published On2026/07/01
• JournalInternational Journal of Advanced Multidisciplinary Application (IJAMA)
• ISSN No.3048-9350

Abstract

Post-harvest fungal contamination of stored wheat grain, predominantly by Fusarium, Aspergillus, and Penicillium species, causes substantial economic loss through grain spoilage and mycotoxin-related batch rejection, yet conventional quality control relies on visual inspection and laboratory mycotoxin assays that detect contamination only after it has progressed to a visibly or chemically advanced stage. This study addresses the resulting food-safety and agricultural-engineering gap by developing and validating a deep learning framework that processes hyperspectral imaging (HSI) data to detect fungal infection in wheat kernels substantially earlier than visible symptoms or mycotoxin thresholds would permit, integrating optical sensing engineering, food microbiology, and computer vision methods within a single non-destructive screening system.
Wheat kernels (cultivar HD-3086) were inoculated with Fusarium graminearum, Aspergillus flavus, and Penicillium spp. under controlled storage conditions (28°C, 85% relative humidity) and imaged daily across a push-broom hyperspectral imaging system spanning 400-1000 nm (240 spectral bands) over a 14-day storage trial, with parallel deoxynivalenol (DON) mycotoxin quantification by enzyme-linked immunosorbent assay providing ground-truth contamination severity labels. A three-dimensional convolutional neural network (3D-CNN) jointly exploiting spectral and spatial kernel-surface information was trained for four-class infection severity classification (healthy, early-stage, moderate, severe) and benchmarked against a 2D-CNN (spectral features only), partial least squares discriminant analysis (PLS-DA), and a support vector machine, using a dataset of 5,600 individually imaged and ground-truth-labelled kernels.
The 3D-CNN achieved 95.7% four-class classification accuracy and area-under-curve values of 0.911-0.971 for binary infection detection across the three fungal species, exceeding the 2D-CNN (91.8% accuracy) and classical baselines (82.3-84.6% accuracy). Critically, the HSI-CNN system detected infection with 50% cumulative detection probability by storage day 3.2, compared to day 8.5 for simulated visual inspection thresholds, a 5.5-day earlier detection window that precedes the steep phase of deoxynivalenol accumulation entirely. Saliency-based wavelength importance analysis identified the 680 nm chlorophyll-absorption band and the 700 nm red-edge transition as the two most discriminative spectral features, consistent with established plant-pathology mechanisms linking fungal infection to pigment degradation. Economic analysis indicates that HSI-CNN continuous screening reduces annualised per-tonne storage losses by 71.9% relative to no monitoring and by 56.8% relative to periodic visual inspection. These findings demonstrate that coupling optical sensing engineering with deep learning-based spectral-spatial analysis provides a viable non-destructive early-warning system for grain storage quality management.

Keywords: hyperspectral imaging, deep learning, convolutional neural network, fungal contamination, wheat grain, mycotoxin, food safety, post-harvest, non-destructive sensing, precision agriculture

References

1. [1] Barbedo, J. G. A. (2019). A review on the use of unmanned aerial vehicles and imaging sensors for monitoring and assessing plant stresses. Drones, 3(2), 40.
2. [2] Femenias, A., Gatius, F., Ramos, A. J., Sanchis, V., & Marin, S. (2020). Hyperspectral imaging for the classification of individual cereal kernels according to fungal and mycotoxin contamination: A review. Food Research International, 137, 109368.
3. [3] Gowen, A. A., O'Donnell, C. P., Cullen, P. J., Downey, G., & Frias, J. M. (2007). Hyperspectral imaging - an emerging process analytical tool for food quality and safety control. Trends in Food Science & Technology, 18(12), 590-598.
4. [4] Jia, B., Wang, W., Ni, X., Lawrence, K. C., Zhuang, H., Yoon, S. C., & Gao, Z. (2020). Essential processing methods of hyperspectral images of agricultural and food products. Chemometrics and Intelligent Laboratory Systems, 198, 103936.
5. [5] Liang, P. S., Haff, R. P., Hua, S. S. T., Munyaneza, J. E., Mustafa, T., & Sarreal, S. B. L. (2018). Nondestructive detection of fungal infection in wheat kernels using NIR hyperspectral imaging. Food Control, 89, 80-87.
6. [6] Mahesh, S., Manickavasagan, A., Jayas, D. S., Paliwal, J., & White, N. D. G. (2008). Feasibility of near-infrared hyperspectral imaging to differentiate Canadian wheat classes. Biosystems Engineering, 101(1), 50-57.
7. [7] Senthilkumar, T., Jayas, D. S., & White, N. D. G. (2016). Detection of fungal infection and Ochratoxin A contamination in stored wheat using near-infrared hyperspectral imaging. Journal of Stored Products Research, 65, 30-39.
8. [8] Shen, F., Huang, S., Dai, M., Wu, J., Cao, Y., Fan, Y., Su, Y., Xu, J., & Yan, J. (2022). Hyperspectral imaging combined with deep learning for early detection of fungal infection in wheat. Computers and Electronics in Agriculture, 198, 107071.
9. [9] Su, W. H., & Sun, D. W. (2018). Fourier transform infrared and Raman and hyperspectral imaging techniques for quality determinations of powdery foods: A review. Comprehensive Reviews in Food Science and Food Safety, 17(1), 104-122.
10. [10] Vermeulen, P., Ebrahimi, T., Fernandez Pierna, J. A., Dardenne, P., & Baeten, V. (2017). Authenticity and traceability of grains by NIR hyperspectral imaging. Cereal Foods World, 62(3), 90-95.
11. [11] Wang, L., Pu, H., & Sun, D. W. (2021). Estimation of deoxynivalenol concentration in wheat kernels using hyperspectral imaging and deep learning. Food Chemistry, 360, 130044.
12. [12] Zhang, N., Yang, G., Pan, Y., Yang, X., Chen, L., & Zhao, C. (2020). A review of advanced technologies and development for hyperspectral-based plant disease detection in the past three decades. Remote Sensing, 12(19), 3188

Explore Our Related Journals

Looking for the right journal for your next manuscript? Explore our international peer-reviewed journals covering engineering, management, computer science, artificial intelligence and multidisciplinary research.