Semi-Autonomous Water Drone for Trash Collection with Computer Vision

This project focuses on environmental sustainability through the development of a semi-autonomous water drone designed to clean trash from rivers and lakes. The drone is equipped with a RealSense D435i camera for computer vision-based trash detection and navigation, powered by a Raspberry Pi 5 for efficient processing. The body was designed in Fusion 360 and 3D-printed for lightweight and durable performance, enabling the drone to operate for 4-5 hours on a single charge.

Highlights:

- Computer Vision: Utilized RealSense D435i for trash detection and obstacle avoidance.
- Sustainable Design: Engineered using Fusion 360 for a durable, water-resistant, and eco-friendly structure.
- Efficient Operation: Operates autonomously for up to 5 hours on a single charge.
- Recognition: Finalist in the Kazenergy Eco-Shell competition, showcasing the project’s innovation and impact.

3D-Printed Corpus:

Bottle Detection:

Teleop test:

The water drone demonstrates the potential of robotics in addressing environmental challenges by efficiently collecting trash from water bodies, making a significant contribution to ecological conservation.