Back to Projects
University Project

LeafScan - Plant Disease Detection and Care Advisory System

ReactTypeScriptFastAPITensorFlowReact NativeMongoDBSupabase
Source Code Docs Live Demo
LeafScan - Plant Disease Detection and Care Advisory System

Overview

Built with React, TypeScript, FastAPI, React Native and TensorFlow. Features include image upload, disease detection using a CNN model, and care advisory based on detected diseases, Weather data integration for specific care advices and a community forum.

Key Features

  • Image Upload for Disease Detection
  • CNN Model for Accurate Disease Classification
  • Care Advisory Based on Detected Diseases
  • Weather Data Integration for Specific Care Advices
  • Community Forum for User Interaction and Knowledge Sharing

Architecture & Decisions

  • CNN Model SelectionCNN works best for image classification tasks, providing high accuracy in disease detection from leaf images.
  • FastAPI for BackendChose FastAPI for its high performance and ease of use with Python, allowing for rapid development of RESTful APIs.
  • React Native for Mobile AppLogic reuse and cross platform development
  • MongoDB for primary databaseData is unstructured and needs flexibility
  • Microservices ArchitectureImplemented microservices architecture to separate the AI model service from the main application, allowing for independent scaling and maintenance of each distinct functionality.
  • OpenSearch for ForumEasiEfficienter to implement full text search capabilities and future extension for logging

Challenges & Solutions

Challenge / Problem

Limited Time

Solution / Implementation

Used Agile methodology with 2 week sprints to prioritize features and deliver a functional system within the project timeline.

Challenge / Problem

Deployment Constraints

Solution / Implementation

Deployed the AI model as a separate microservice using FastAPI, allowing for independent scaling and maintenance. Used Docker containers to simplify deployment and ensure consistency across environments.

Challenge / Problem

Insufficient Data for Certain Diseases

Solution / Implementation

Used Data Augmentation techniques to artificially expand the dataset, improving model robustness and accuracy for underrepresented diseases.

Project Timeline

Week 1

Requirement Gathering & Planning

Week 2

Architecture Design & Prototyping

Week 3-8

Development

Week 9-10

Testing

Week 11-12

Deployment

Gallery

Architecture Diagram

Architecture Diagram

A high-level overview of the LeafScan system architecture and corresponding technologies used

Homepage

Homepage

Clean and responsive landing page with user-friendly navigation

Dashboard

Dashboard

A User Friendly Dashboard for Easy Navigation

Community Forum

Community Forum

Community Forum for User Interaction and Knowledge Sharing