Reverse Vending Machine

Automated Recycling System with Vision-Based Identification

Automated Reverse Vending Machine System Concept

This project focused on the conceptual design and development of an automated reverse vending machine system for plastic bottle recycling. The system was designed to identify, validate, and sort recyclable bottles through a combination of computer vision, weight verification, embedded processing, and automated actuation.

The project was conducted as a research project at the University of Sri Jayewardenepura in collaboration with OREL Corporation (Pvt.) Ltd.

Blender-based simulation of the proposed automated reverse vending machine concept.

System Architecture and Workflow

The proposed system was designed to automate the bottle recycling process through an integrated perception, decision-making, and sorting pipeline.

The workflow incorporated bottle identification, database verification, weight-based validation, and actuator-controlled sorting into separate collection bins.

System workflow developed for the proposed automated reverse vending machine, outlining bottle verification, weight validation, and automated sorting logic.

Digital Prototype Development

A complete digital prototype was developed using Blender to evaluate the mechanical layout and demonstrate the automated bottle sorting concept before physical implementation.

The simulation included the bottle insertion process and the movement of bottles through the proposed sorting mechanism.

Blender-based digital prototype illustrating the mechanical layout and bottle sorting mechanism from multiple viewpoints.

Vision-Based Bottle Identification

Initial experiments investigated camera-based identification approaches for automated bottle recognition.

A webcam-based perception system was developed to investigate bottle recognition using barcode identification as part of the proposed sensing pipeline.

Initial computer vision experiments for automated bottle identification and barcode recognition.

Embedded System Design

The proposed system architecture combined multiple hardware components to enable automated operation, including camera-based identification, weight verification, embedded processing, and actuator-based sorting.

The planned implementation included:

  • Camera-based bottle identification
  • Weight measurement using a load cell
  • Embedded processing using a Raspberry Pi
  • Servo-controlled sorting mechanism

The project demonstrated the design of a complete automated recycling workflow, from user interaction and perception to decision making and physical sorting.