AI-Based Arecanut Plant Health Detection and Management System Using a Custom Convolutional Neural Network
Arecanut is an important plantation crop and its productivity can be affected by diseases, pests, and other visible plant-health problems. Early identification of such conditions can support timely inspection and management. This work describes an image-based Arecanut Plant Health Detection and Management System developed using a custom Convolutional Neural Network (CNN) and a local Flask web application. The proposed model is configured to classify images into ten Arecanut-related health, disease, and pest categories. The training pipeline uses 224 × 224 RGB input images, data augmentation, class-weight balancing, batch normalization, four convolutional blocks, global average pooling, a dense layer, dropout, and softmax classification. The trained model is integrated with a practical web application that displays prediction results together with category-specific severity, treatment information, and recommended actions. The completed application additionally provides English and Kannada interfaces, SQLite-based prediction history, disease progress tracking, a Smart Treatment Planner, CSV history download, and Clear History functionality. The system was also tested with the Internet connection disabled, demonstrating that the core model and application workflow can operate using locally stored resources. The project therefore combines image-based plant-health classification with a management-oriented application intended for practical and accessible use.
Authors
- Boodalu Priya
- Divya S
- Ganashree M S
- Priyanka G.V
- Chaithra M H
Publication Details
- Journal
- Zenodo (CERN European Organization for Nuclear Research)
- Published
- 2026-11-19
- DOI
- https://doi.org/10.5281/zenodo.24360439
- Primary Topic
- Smart Agriculture and AI
- Type
- article
- Field-Weighted Citation Impact
- 0.00