Plastic Waste Detection on Beaches Using CNN and YOLO-Based Image & Video Analytics
Summary
No summary yet. VIRAT can read this paper and write a plain-language summary of its aim, methods, key findings and limitations.
Summarise with VIRAT Free to try with your Science Catalogs account.Abstract
Beach litter of plastics is an escalating conservation challenge that harms ocean- and coast-dwelling wildlife, degrades coasts, and threatens human communities reliant on healthy tourism. In this project, our aim is to develop a useable deep learning solution that will find and map plastic waste automatically from the drone footage or fixed camera images. Our method uses a combination of CNNs, which can spot features such as the shape and color of debris, along with YOLO, to help us pinpoint and outline each piece of plastic accurately. We collected diverse kinds of images -some from available open-source datasets (e.g TACO, Ocean Cleanup, TrashNet). own photos and videos, and we also took care to how important it was for the system to work in various weather conditions and lighting. To test how well our system worked, we examined its accuracy, precision, recall and OK-state F1-score in order to show that we can separate the typed of plastic waste effectively. Our results indicate that combination of CNN and YOLO is beneficial for monitoring beach cleaniness. In the future, we'd like to add features to track waste in videos, segmenting more precise shape of it and connecting our findings with maps that would make cleanups planning easier as well as support the larger goal of committing (tm) volunteering. protecting coastal areas and our oceans. This work contributes to the United Nations Sustainable Development Goal 14: Life Below Water, which aims at conserving and sustainably using the oceans, seas, and marine resources.
Fields of Study & Subjects
Microplastics and Plastic Pollution
Advanced Neural Network Applications
Remote-Sensing Image Classification
Publication details
- Journal / source
- International Conference on Intelligent Systems for Communication, IoT and Security (ICISCoIS)
- Pages
- 1–7
- Publication date
- 8 January 2026
- Language
- EN