AI-Powered Plant Disease Detection System Using CNN and Flask — Final Year Project with Source Code

Core Features of the Plant Disease Detection System

Deep Learning Inference with PyTorch CNN

At the heart of PlantAI is a trained Convolutional Neural Network model built using PyTorch. The model classifies plant leaf images into one of 39 output classes with high accuracy. Each prediction comes with a confidence score, allowing the end-user to understand how certain the AI is about its diagnosis. This makes it an excellent demonstration of applied deep learning and computer vision concepts, which are central to most AI final year project evaluations.

39 Disease Classes Across 13 Crops

The classifier covers a wide and realistic range of crop diseases. Supported crops include Apple, Blueberry, Cherry, Corn, Grape, Orange, Peach, Pepper, Potato, Raspberry, Soybean, Squash, Strawberry, and Tomato. Disease conditions range from bacterial spot and powdery mildew to late blight, leaf scorch, and mosaic virus, along with a dedicated class for healthy leaves and background without plant material. This breadth makes the project suitable for agricultural technology demonstrations and research presentations.

Image Upload and Mobile Camera Capture

Users can diagnose plant diseases either by uploading a photograph from their device or by using the mobile camera directly within the browser. This dual input mode makes PlantAI practical in real field conditions, which is a compelling selling point for academic presentations and viva evaluations.

CSV-Backed Disease Information and Treatment Guidance

Once a prediction is made, the system fetches disease descriptions, prevention steps, and treatment guidance from a structured CSV file. This approach demonstrates clean data management and separation of concerns in software design, which evaluators appreciate in a well-architected final year project with source code.

Supplement and Fertilizer Marketplace Integration

PlantAI includes a built-in marketplace tab that recommends relevant fertilizers and agricultural supplements based on the detected disease. Product data is maintained in a separate CSV file, and the interface links out to purchase pages. This feature demonstrates practical integration of e-commerce logic within an AI application.

Weather-Based Disease Risk Analysis

The application connects to the Open-Meteo weather API to perform a rule-based disease risk analysis based on current and forecast weather conditions. This adds a real-world environmental dimension to the project and showcases students ability to integrate third-party APIs within a machine learning application.

Groq-Powered AI Plant Care Assistant

A built-in conversational assistant powered by the Groq API allows users to ask plant care questions using typed text or browser voice input. This generative AI integration elevates PlantAI well beyond a standard image classifier, making it one of the more feature-complete Python final year projects available for academic submission.

Stripe and COD Payment Integration

The supplement storefront supports Stripe Checkout for online payments and a COD order-capture path for offline transactions. Including real payment logic in a final year project demonstrates practical full-stack development skills that go beyond academic toy examples.

PWA Support and Responsive Design

PlantAI ships with a PWA manifest and service worker configuration, making it installable on mobile devices. The frontend is built with Bootstrap 5 and custom CSS, ensuring a clean, responsive experience across screen sizes.