Enhancing Jute Crop Management with Hybrid EfficientNetB7 for Disease Prediction

Abstract

Abstract Jute is one of the most important cash crops in Bangladesh and plays a vital role in the country's economy. However, the yield and quality of jute are often affected by various plant diseases, which are difficult to detect manually at an early stage. Early identification is essential to prevent large-scale damage and ensure sustainable production. This study proposes an automated jute disease detection system using image processing and deep learning techniques. Several models, including a Custom Convolutional Neural Network (CNN), VGG16, DenseNet121, and a Hybrid EfficientNetB7 architecture, were implemented and compared. The proposed Hybrid EfficientNetB7 model achieved the highest classification accuracy of 99.7%, outperforming the other models. The system effectively distinguishes between healthy and diseased jute leaves and stems, providing a reliable tool for early disease detection. This research demonstrates that deep learning-based ensemble and transfer learning methods can significantly improve the accuracy and reliability of plant disease detection systems. The findings can contribute to the development of intelligent agricultural tools, empowering farmers with faster and more accurate disease diagnosis for better crop management and yield improvement.

Code and data availability

植物フェノタイピング解析を再現する公開資産であることを、入力本文と直接リンクから確認できなかったため保留しました。

No evidence-backed public reproduction asset is currently recorded.