Automatic Plastic Waste Segregation

Abstract

An efficient plastic waste segregation system that integrates computer vision and machine learning to achieve accurate and real-time sorting of plastic materials such as PET, HDPE, and PVC. A Raspberry Pi equipped with a camera module captures images of plastic waste transported on a conveyer belt, and a deep learning–based classification model, trained using a labeled dataset divided into training, validation, and testing sets, identifies the plastic type based on visual features. The predicted classification is communicated to an Arduino UNO via serial communication, which performs real-time control of a DC motor and a servo-driven mechanical diverter through motor driver circuit to physically segregate the waste into designated bins. The system is powered by a hybrid energy supply, where solar energy is utilized to operate the entire system during peak sunlight hours, while grid power is automatically used during off-peak or low-sunlight conditions to ensure uninterrupted operation. By reducing manual intervention and improving segregation accuracy, the system not only offers a scalable and cost-effective solution for smart plastic waste management but also promotes sustainability through the effective use of renewable energy resources.

Introduction

The increasing production of plastic waste has become a major global environmental problem, as it contributes significantly to land and ocean pollution and harms ecosystems. Traditional manual waste sorting is time-consuming, error-prone, inefficient, and unsafe for workers. To address these challenges, the study proposes an automated plastic waste segregation system using computer vision, artificial intelligence, and embedded systems.

The system uses a Raspberry Pi as the main processing unit and the YOLOv8 deep learning algorithm to detect and classify plastic types such as PET, HDPE, and PVC from images captured by a camera mounted above a conveyor belt. After classification, the system automatically controls servo motors to divert plastics into appropriate bins. The conveyor belt system ensures continuous movement, while control signals are managed through an Arduino UNO, motor drivers, and logical control mechanisms.

To enhance sustainability, the system is powered by a solar photovoltaic energy source with battery storage and voltage regulation, reducing reliance on grid electricity.

The system was tested using MATLAB/Simulink simulation to validate conveyor control and servo motor operations before hardware implementation. Simulation results showed proper coordination between detection and sorting mechanisms. In hardware implementation, the system successfully demonstrated real-time detection and automated sorting.

Overall, the proposed system improves accuracy, speed, safety, and environmental sustainability in plastic waste management while reducing dependence on manual labor and conventional energy sources.

Conclusion

Automated plastic waste segregation using computer vision and embedded control techniques provides an effective approach for improving waste management efficiency. The developed system integrates a Raspberry Pi processing unit, Pi Camera module, Arduino UNO controller, and electromechanical sorting mechanism to automatically detect and separate plastic waste materials. The YOLOv8 deep learning model enables real-time object detection and classification of different plastic categories such as PET, HDPE, and PVC. Images captured from the conveyor belt are processed by the Raspberry Pi, and the classification results are transmitted to the Arduino UNO, which controls the servo motor and conveyor motor for the sorting operation. Continuous movement of waste materials through the conveyor mechanism allows efficient automated segregation without manual intervention. A renewable energy power system supported by a solar photovoltaic module and battery storage improves system sustainability, while a buck converter regulates the battery voltage to a stable 5 V supply required for the Arduino and other low-voltage electronic components. Reliable detection performance achieved by the YOLOv8 model demonstrates the suitability of deep learning–based approaches for real-time waste segregation applications. The overall system provides a compact, cost-effective, and scalable solution that can reduce manual labor, improve sorting accuracy, and contribute to environmentally sustainable plastic waste management.

References

[1] S. Munira, N. Paul, M. A. Alam, M. M. Rahman, and M. S. Kaiser, “Turning trash into treasure: Developing an intelligent bin for plastic bottle recycling,” Journal of Social Computing, vol. 5, no. 1, pp. 1–14, 2024. [2] R. Procházka, J. Valí?ek, M. Harni?árová, M. Kušnerová, H. Tozan, C. Borzan, M. Kadnár, Z. Palková, R. Gálik, and K. Slamová, “Collection of plastic packaging of various types: Sorting of fractions of plastic waste using both automated and manual modes,” IEEE Access, vol. 12, pp. 44244–44261, 2024. [3] S. Hamzah, N. N. A. Mangshor, M. S. Kamarudin, S. Ibrahim, and A. F. A. Fadzil, “Recyclable waste detection using You Only Once (YOLOv8): A preliminary study,” in 2024 IEEE 22nd Student Conference on Research and Development (SCOReD), pp. 488–493, 2024. [4] K. Patel, V. Patel, V. Prajapati, D. Chauhan, A. Haji, and S. Degadwala, “Safety helmet detection using YOLO v8,” in 2023 3rd International Conference on Pervasive Computing and Social Networking (ICPCSN), pp. 22–26, 2023. [5] A. Jain, A. Pasha H, V. Balaji, R. Rohitha and K. K. B M, \"Design and Development of Solar Powered Automatic Grain Dryer for Storage,\" 2019 IEEE International Conference on Distributed Computing, VLSI, Electrical Circuits and Robotics (DISCOVER), Manipal, India, 2019, pp. 1-5, doi: 10.1109/DISCOVER47552.2019.9008041. [6] B. M. Kiran Kumar, M. Rachana, H. G. Rakshitha, J. Dhanush and K. C. Kallajjanavar, \"An Innovative Simplistic Solar Powered Low-Cost Dry Waste Segregator Robot,\" 2023 International Conference on the Confluence of Advancements in Robotics, Vision and Interdisciplinary Technology Management (IC-RVITM), Bangalore, India, 2023, pp. 1-5, doi: 10.1109/IC-RVITM60032.2023.10434953. [7] I. N. Mhlongo, G. U. Nnachi, A. F. Nnachi, and A. T. Adesola, “Modelling and simulation of conveyor belt for energy efficiency studies,” in 2020 IEEE PES/IAS PowerAfrica, pp. 1–5, 2020. [8] A. Kulikov, V. Kaverin, and A. Mosavi, “Simulation of an electric conveyor drive using Simulink Matlab,” in 2024 IEEE 2nd World Symposium on Applied Machine Intelligence and Informatics (SAMI), pp. 000513–000518, 2024.

Copyright

Copyright © 2026 Joval Jiju, Revathi V R, Akshai J Mundakkal, Janasudhan K, Prof. Jeena Joy, Dr. Sera Mathew. This is an open access article distributed under the Creative Commons Attribution License, which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited.