`
Introduction
RUTrash is a sustainable initiative developed with the goal of addressing environmental conservation challenges through innovative technology solutions. This project encompasses various functionalities, including trash type detection, carbon footprint calculation, and a climate expert chatbot powered by Cloudflare Workers AI API. In this detailed write-up, we'll delve into the development process, technologies used, challenges faced, accomplishments, and lessons learned throughout the creation of RUTrash.
Inspiration and Objectives
The inspiration for RUTrash stems from the urgent need to tackle environmental issues, particularly in waste management and carbon emissions. The project aligns with the mission of the Rutgers Office of Climate Action, aiming to empower individuals to make environmentally conscious decisions in their daily lives. The key objectives of RUTrash are:
- Trash Type Detection: Enable users to identify and properly dispose of various types of waste through image analysis.
- Carbon Footprint Calculation: Provide users with insights into their environmental impact based on daily habits and activities.
- Climate Expert Chatbot: Engage users in conversations about climate sustainability and offer guidance on contributing to environmental protection.
Development Process
Training MobileNetV2 on Kaggle's Garbage Classification Dataset
The development journey began with selecting a suitable dataset for garbage classification. The Kaggle dataset, known for its diversity and quality, was chosen. The dataset underwent preprocessing, including image resizing, normalization, and augmentation, to prepare it for training. MobileNetV2, a lightweight and efficient CNN architecture, was selected and trained using PyTorch and Torchvision libraries. Validation and fine-tuning were crucial steps to ensure the model's accuracy and generalization.
Implementing the Application with Flask
Flask, a Python web framework, was instrumental in building the backend of RUTrash. Flask facilitated image upload and processing for trash type detection, handled user inputs for carbon footprint calculation, and integrated seamlessly with Cloudflare Workers AI API for the climate expert chatbot. The Flask setup involved defining routes, endpoints, and backend processing logic to deliver a user-friendly and responsive application.
Integration with Cloudflare Workers AI API
Cloudflare Workers AI API played a pivotal role in powering the Climate Expert Chatbot feature. Integration with the API involved understanding API documentation, handling API requests and responses, and ensuring smooth communication between Flask backend and the external AI service. This integration enriched the user experience by providing personalized guidance on climate sustainability.
Functionalities of RUTrash
-
Trash Type Detection
Description: Users can upload waste images for classification.
Functionality: MobileNetV2 analyzes uploaded images to identify the type of trash.
Response: Suggestions for appropriate disposal methods are provided based on the classification.
-
Carbon Footprint Calculation
Description: Users input data related to daily habits impacting the environment.
Functionality: Flask processes user inputs to calculate carbon footprint.
Response: Users receive insights into their environmental impact, fostering awareness and sustainable practices.
-
Climate Expert Chatbot
Description: AI-powered chatbot engages users in climate sustainability conversations.
Functionality: Cloudflare Workers AI API powers the chatbot, offering personalized guidance.
Response: Users receive advice on contributing to environmental protection and adopting eco-friendly practices.
Challenges and Solutions
Model Training Complexity
Training MobileNetV2 for garbage classification within a limited timeframe posed challenges. Careful dataset selection, preprocessing, and model optimization strategies were employed to overcome these challenges.
API Integration
Integrating Cloudflare Workers AI API required understanding API documentation and ensuring seamless communication. Close collaboration and testing were essential for successful integration.
Accomplishments and Learnings
Accomplishments
- Successful implementation of all planned functionalities, including trash type detection, carbon footprint calculation, and the Climate Expert Chatbot.
- Going beyond initial goals by adding extra features, showcasing dedication and problem-solving skills.
Learnings
- Application of CNNs, Flask, HTML, and CSS for developing a fully functional and impactful application.
- Insights into AI integration, web development, and environmental sustainability, enriching the learning experience.
Future Enhancements
- Transforming the application into a versatile platform accessible across various devices for wider reach and usability.
- Enhancing the UI/UX of the webpage to improve user experience and engagement.
- Implementing a more robust CNN model for image classification to further enhance accuracy and reliability.
Conclusion
RUTrash exemplifies a comprehensive approach to addressing environmental conservation through technology. The integration of AI, Flask development, and API integration has resulted in a powerful platform that empowers users to make informed decisions and contribute to a more sustainable future.