• Yugandhar Reddy
• Uday Kiran
• Chakradhar Reddy

The environmental impact of underwater garbage poses significant challenges for marine ecosystems and requires effective monitoring and management strategies. In this research, we present a novel approach to address the critical issue of real time underwater garbage detection using state-of-the-art object detection and image segmentation models based on the You Only Look Once (YOLO) architecture. Our study aims to provide a reliable and efficient system for identifying and classifying different types of underwater garbage, contributing to environmental preservation efforts and enabling better marine habitat monitoring. To achieve accurate and real-time garbage detection, we evaluate several YOLO based models, including YOLOv5s and Spatial Attention Module (SAM). These models are known for their efficiency in processing images and detecting multiple objects simultaneously. To facilitate this research, we develop a cumulative, self annotated dataset that includes various underwater garbage instances, such as trash, plastic, and underwater debris, as well as other marine entities like fish and flora & fauna.We conduct a comprehensive comparative analysis of the YOLO-based models, considering their performance metrics such as accuracy, precision, recall, and processing speed. The evaluation results offer valuable insights into the strengths and weaknesses of each model, enabling us to identify the most effective one for underwater garbage detection. Furthermore, we emphasize the importance of efficient implementation to address the challenges posed by limited storage space and computing capabilities of underwater mobile devices. Our system is designed to enable real-time detection and classification on resource-constrained devices, ensuring practical applicability in underwater monitoring missions. The proposed real-time underwater garbage detection system has the potential to significantly impact environmental preservation efforts and marine conservation initiatives. By providing a reliable method for identifying and classifying underwater garbage, it aids in better understanding the distribution and abundance of marine litter and its potential effects on marine life. In conclusion, this research project contributes to the development of a robust and efficient real-time underwater garbage detection system by leveraging YOLO-based object detection and image segmentation models. The study's outcomes can inform decision-makers and environmental agencies, helping them devise effective strategies for mitigating the impact of underwater garbage on marine ecosystems.

Underwater Garbage Detection, YOLO-based Object Detection, Image Segmentation, Marine Conservation, Real-time Monitoring, Environmental Preservation

"Real-time Underwater Garbage Detection with YOLO-based Object Detection and Image Segmentation Models ", International Journal of Emerging Technologies and Innovative Research (www.jetir.org), ISSN:2349-5162, Vol.11, Issue 3, page no.g68-g76, March-2024, Available :http://www.jetir.org/papers/JETIR2403609.pdf

"Real-time Underwater Garbage Detection with YOLO-based Object Detection and Image Segmentation Models ", International Journal of Emerging Technologies and Innovative Research (www.jetir.org | UGC and issn Approved), ISSN:2349-5162, Vol.11, Issue 3, page no. ppg68-g76, March-2024, Available at : http://www.jetir.org/papers/JETIR2403609.pdf