DEVELOPMENT OF A PLANT DISEASE DETECTION SYSTEM USING CONVOLUTIONAL NEURAL NETWORKS

In the current agricultural environment, effectively managing crop health is essential due to growing challenges from disease outbreaks (Pawlak & Kołodziejczak, 2019). Traditional methods can be labor-intensive and often require expert help, making it difficult for small-scale farmers and researchers. The primary aim of this project is to develop a Plant Disease Identification system that uses image processing to deliver accurate and timely disease diagnoses. The project uses the Waterfall SDLC framework, including requirements analysis, design, implementation, and evaluation. A CNN model was developed by training it on plant disease images with TensorFlow/Keras in Google Colab. The system was built with HTML, CSS, JavaScript and also connects to a flask backend for interaction with the trained model. The AI model achieved a classification accuracy of 984.2% on test data, demonstrating robust performance in identifying various crop diseases. The web interface successfully allows users to upload images and receive immediate diagnostic results. Usability testing confirmed that the system is intuitive, responsive, and suitable for both technical and non-technical users. The Ag-Cure platform proves the viability of AI in enhancing agricultural disease diagnosis. It provides a scalable, costeffective, and user-accessible solution that empowers farmers to manage crop health efficiently. Future improvements may include expanding the disease dataset, supporting offline usage via Progressive Web Apps (PWA), and integrating regional language support and IoT-based monitoring systems.