Date of Publication :15th June 2024

Abstract:This research paper presents a new method for plant disease diagnosis using neural network architectures, specifically convolutional neural networks (CNN). First, we will highlight the importance of plant disease testing and the problems associated with traditional methods. Our literature review highlights recent advances in neural network architectures (particularly CNNs) for plant disease detection, highlighting the growing interest and progress in using machine learning for this task. The methodology section describes three methods, including data selection, model design, training methods, and evaluation methods. We present the results of a successful trial on publicly available data that demonstrates the validity of our proposal. Additionally, we integrate the trained CNN model with the Python flask framework to develop a user-friendly website for real-time virus detection. In this project, the model we have presented has achieved 98.9% test accuracy and 98.7% verification accuracy with an impressive train accuracy of 96.7% were achieved. In conclusion, we summarise our findings and highlight the importance of the contribution of our research to agricultural science and technology. Our proposed system has the potential to complement the discovery process, promote permaculture, and reduce crop losses due to plant diseases.

Reference :