🪴 Planteria: AI‑Powered Plant Disease Detection & Botanist Assistant

---

🌿 Project Overview

Welcome to Planteria, an innovative AI-driven solution designed to empower gardeners and farmers with instant plant disease diagnosis and expert botanical advice. This project showcases a full‑stack approach to solving real-world agricultural challenges by integrating cutting-edge Deep Learning for computer vision with a conversational AI interface. It also lays the groundwork for potential e‑commerce integration.

Problem: Plant diseases cause significant crop losses, affecting food security and farmer livelihoods. Traditional diagnosis is often slow, requires expert intervention, and can be subjective.

Solution: Planteria offers a rapid, accessible, and accurate system for plant disease identification from images, paired with expert remedies. It also includes a botanist chatbot for broader plant care advice—acting as a virtual agricultural expert. The system is built for scalability and user accessibility.

---

🔑 Key Features

• 📷 Plant Disease Detection
  Upload a plant leaf image to instantly identify potential diseases using a pre-trained Convolutional Neural Network.

• 💊 Automated Remedy Suggestions
  Receive generic, actionable remedies tailored to the detected disease, integrated directly into the prediction output.

• 💬 Botanist AI Chatbot
  Talk to a specialized AI assistant for plant care advice, disease prevention tips, and gardening knowledge. It maintains conversation history and stays strictly on-topic.

• 🌐 RESTful API
  All functionalities (disease prediction, chatbot interaction) are exposed via a robust Flask API, enabling easy integration with future web or mobile apps.

• 🛒 E-commerce Foundation (Conceptual/Future)
  Structure supports adding product listings for remedies, tools, or seeds—paving the way for a full app ecosystem.

---

🚀 Live Demo & Visual Showcase

Click the image above to watch a detailed video demo showing Planteria's features and technical flow.

Live API Endpoint: https://your-hosted-api-url.com
Live Web/Mobile App: [Insert frontend link here]

---

📱 Screenshots

• Upload Interface: Image of a user uploading a plant leaf for diagnosis.
• Diagnosis Result: Screenshot showing predicted disease, confidence score, and remedy.
• Botanist Chatbot: Visual demo of a natural conversation with the AI assistant.

(Replace these placeholders with your actual UI screenshots)

---

📐 System Architecture & Workflow

High-Level System Workflow

Flowchart illustrating how a user's request (image upload or chat) flows through the frontend, Flask API, AI models (CNN and LLM), the remedy database, and returns a response. (Also shows conceptual e-commerce support.)
Save as assets/diagrams/system_workflow.png.

Database Design (ERD)

Entity‑Relationship Diagram outlining core tables (Plants, Diseases, Remedies, Users, Products, Orders) and their relationships—demonstrating future e-commerce readiness.
Save as assets/diagrams/erd_diagram.png.

Data Flow for AI Prediction

Diagram detailing steps taken when processing an uploaded image: preprocessing ➝ model inference ➝ remedy lookup ➝ API response.
Save as assets/diagrams/data_flow_ai.png.

---

🛠️ Technologies Used

Category	Technology	Logo	Justification
Backend/API	Python, Flask, Flask-CORS, requests	🔧	Flask serves RESTful endpoints; Cross‑Origin and HTTP support
AI / ML	TensorFlow, Keras, MobileNetV2, Google Gemini 2.0 Flash	🤖	Enables image classification and chatbot capabilities
Libraries	numpy, pandas, matplotlib, seaborn, opencv-python, scikit-learn	📚	Core data handling, visualization, image processing
Development	PyCharm, Jupyter Notebook, Git / GitHub	💼	For efficient coding, experimentation, and version control

Action: add small logos to assets/tech_logos/.

---

🧠 AI / ML Technical Deep Dive

1. Plant Disease Detection

• Objective: Classify plant leaf images as healthy or diseased.
• Dataset: PlantVillage (thousands of labeled images).
• Approach:
  • Transfer Learning: Use MobileNetV2 pre-trained on ImageNet.
  • Data Augmentation: Rotations, zooms, flips, shears to improve robustness.
  • Fine-tuning: Customize top layers; freeze base layers initially.
• Training Plots:
  
  Your analysis here.
• Evaluation:
  • Classification report (precision, recall, F1).
  • Confusion matrix insights.

2. Botanist AI Chatbot

• Objective: Provide contextual plant care advice.
• Approach:
  • Use Google Gemini 2.0 Flash via API.
  • Prompt engineering to establish AI persona and boundaries.
  • Handles off-topic politely and manages conversation context.
  • Remedy integration via internal disease-to-remedy lookup.

---

💻 Backend API (Flask)

🔁 API Endpoints

1. POST /predict_disease

• Description: Submit leaf image ➝ returns disease + remedy.
• Request: multipart/form-data with image.
• Response Example (JSON):

{
  "disease": "Tomato_Early_blight",
  "confidence": 0.985,
  "remedy": "Remove lower leaves, stake for air circulation, apply fungicides like chlorothalonil or mancozeb."
}