AI-Powered Fruit Disease Detection

This repository contains the complete backend and deployment scripts for an AI-powered application designed to solve a practical agricultural problem: the rapid and accurate detection of fruit diseases. It showcases a modern, event-driven, and cost-effective serverless architecture on AWS.

---

Table of Contents

• Key Features
• Live Demo & Examples
• System Architecture
• Getting Started
  • Prerequisites
  • Deployment
• Usage
• Project Structure
• Future Enhancements

Key Features

• AI-Powered Analysis: Uses Amazon Bedrock with the Claude 3.5 Sonnet model for state-of-the-art image recognition.
• Serverless Backend: No servers to manage. The application runs on-demand, ensuring high availability and cost-efficiency.
• Scalable by Design: Built with API Gateway and Lambda, the system automatically scales to handle any amount of traffic.
• Infrastructure as Code (IaC): All AWS resources are defined in a template.yaml file and deployed automatically with AWS SAM, ensuring consistent and repeatable environments.
• Secure by Default: Follows the Principle of Least Privilege with a fine-grained IAM role for the Lambda function.

Live Demo & Examples

This project includes both a user-friendly web interface for easy demonstration and a command-line script for testing the backend API directly.

1. Web Interface Demo

The primary way to interact with the application is through the simple web UI. A user can upload an image and receive the analysis in seconds.

2. API Testing

The backend API can also be tested directly from the command line after deployment:

Example Request:

./test_api.sh path/to/the/fruit_image.jpg

Example Successful Response::

System Architecture

The application follows a classic serverless, event-driven pattern. The architecture is designed for low latency, high scalability, and minimal operational overhead.

Getting Started

Follow these instructions to deploy the application in your own AWS account.

Prerequisites

• AWS CLI configured with your credentials.
• AWS SAM CLI installed.
• Python 3.9+ installed.
• Docker installed and running.

Deployment

1. Clone the repository:

git clone https://github.com/AYassin01/aws-serverless-fruit-detection.git
cd aws-serverless-fruit-detection

2. Enable Model Access:

• Navigate to the Amazon Bedrock console in your AWS account.
• Go to Model access and ensure you have been granted access to Anthropic's Claude 3.5 Sonnet.

3. Run the deployment script:

• The deploy.sh script automates the entire process, including creating the Boto3 layer, building the application, and deploying the AWS resources.

chmod +x deploy.sh
./deploy.sh

• Follow the prompts from the SAM CLI. This will create a CloudFormation stack and deploy all the necessary resources.

Usage

After a successful deployment, the 'deploy.sh' script will create a 'test_api.sh' script for you.

1. Make the test script executable:

chmod +x test_api.sh

2. Run the script with the path to a fruit image:

./test_api.sh path/to/the/fruit_image.jpg

Project Structure

The repository is organized to separate application code from deployment configuration.

├── images/
├── src/
│   └── lambda_function.py
├── .gitattributes
├── .gitignore
├── create_boto3_layer.sh
├── deploy.sh
├── index.html
├── LICENSE
├── README.md
├── samconfig.toml
├── template.yaml
└── test_api.sh

Future Enhancements

• CI/CD Pipeline: Implement a CI/CD pipeline using GitHub Actions to automatically deploy changes pushed to the main branch.
• Asynchronous Processing: For very large images or videos, refactor the architecture to use S3 event notifications and an SQS queue for asynchronous processing.