Plant Disease Detection using Convolutional Neural Network
Abstract
Agriculture is the most important aspect of human life and also plays a vital role in a country’s economy. Plants affected by diseases require huge effort, time, budget, and harmful sprays for treatment. The first symptoms of disease in plants are difficult to detect from the human eye at an early stage. In this study, fungus disease detection is performed using Convolutional Neural Network (CNN). This study focuses on detecting healthy or unhealthy plant leaves based on the very first fungus symptoms with high accuracy and an automated process. A pre-trained model VGG-16 using transfer learning is adopted, which is further fine-tuned and trained on the target plant’s dataset for its upper layers. Further, a plant disease detection android application is provided for the user to detect fungus disease in plants. The android application uses the picture of the affected leaf to classify healthy and infected leaves. The proposed model showed 99% accuracy, which is better than other state-of-the-art methods.
Copyright (c) 2021 Journal of Information Communication Technologies and Robotic Applications

This work is licensed under a Creative Commons Attribution-NonCommercial 4.0 International License.