From images to insights: Using a convolutional neural network to improve powdery mildew severity detection in mungbean

Pitchakon Papan, Witsarut Chueakhunthod, Chanwit Kaewkasi, Wanploy Jinagool, Akkawat Tharapreuksapong, Teerayoot Girdthai, Kanlayanee Sawangsalee and Piyada Alisha Tantasawat
Additional contact information
Pitchakon Papan: School of Crop Production Technology, Institute of Agricultural Technology, Suranaree University of Technology, Nakhon Ratchasima, Thailand
Witsarut Chueakhunthod: School of Crop Production Technology, Institute of Agricultural Technology, Suranaree University of Technology, Nakhon Ratchasima, Thailand
Chanwit Kaewkasi: School of Computer Engineering, Institute of Engineering, Suranaree University of Technology, Nakhon Ratchasima, Thailand
Wanploy Jinagool: School of Crop Production Technology, Institute of Agricultural Technology, Suranaree University of Technology, Nakhon Ratchasima, Thailand
Akkawat Tharapreuksapong: Center for Scientific and Technological Equipment, Suranaree University of Technology, Nakhon Ratchasima, Thailan
Teerayoot Girdthai: School of Crop Production Technology, Institute of Agricultural Technology, Suranaree University of Technology, Nakhon Ratchasima, Thailand
Kanlayanee Sawangsalee: School of Crop Production Technology, Institute of Agricultural Technology, Suranaree University of Technology, Nakhon Ratchasima, Thailand
Piyada Alisha Tantasawat: School of Crop Production Technology, Institute of Agricultural Technology, Suranaree University of Technology, Nakhon Ratchasima, Thailand

Plant Protection Science, vol. preprint

Abstract: To efficiently identify powdery mildew (PM) severity in mungbean leaves, we developed a Convolutional Neural Network (CNN) approach and validated its effectiveness against human evaluation. We fine-tuned an EfficientNet-B3 pre-trained model, which, in our related studies, performed better than re-implemented Inception V3 models. The CNN was trained on 90% of the images (2 880) for training and 10% (320) for validation, with data augmentation applied using Python and TensorFlow. The model obtained 82.10% and 73.03% as training and validation accuracies after 14 epochs, respectively. Further analysis with an additional 15 datasets revealed PM disease indices ranging from 2.03 (resistance) to 6.45 (high susceptibility). The concordance between AI-predicted and human-assessed PM severity was 74.4% (adjusted R2: 72.4%), with an average root mean squared error (RMSE) of 0.854 and a mean absolute error (MAE) of 0.715, indicating moderate predictive error. Comparison of our developed AI-based application prototype on smartphones with expert evaluations yielded a strong correlation (r = 0.992**, R2 = 0.984), suggesting that this tool can effectively estimate PM severity across mungbean cultivars. The application shows considerable promise, and further optimisation and strategic dissemination efforts will enhance its adoption among farmers.

Keywords: Artificial intelligence; deep learning; EfficientNet-B3; machine learning; Vigna radiata (L.) R. Wilczek; visual observation (search for similar items in EconPapers)
References: Add references at CitEc
Citations:

Downloads: (external link)
http://pps.agriculturejournals.cz/doi/10.17221/15/2025-PPS.html (text/html)
free of charge

Related works:
This item may be available elsewhere in EconPapers: Search for items with the same title.

Export reference: BibTeX RIS (EndNote, ProCite, RefMan) HTML/Text

Persistent link: https://EconPapers.repec.org/RePEc:caa:jnlpps:v:preprint:id:15-2025-pps

DOI: 10.17221/15/2025-PPS

Access Statistics for this article

Plant Protection Science is currently edited by Ing. Eva Karská, (Executive Editor PPS)

More articles in Plant Protection Science from Czech Academy of Agricultural Sciences
Bibliographic data for series maintained by Ivo Andrle ().