Underwater Debris Detection Using Machine Learning

Underwater particles pose widespread environmental challenges, requiring strong detection for effective monitoring and cleanup. This research introduces a system mastering-based totally technique for underwater particles detection, utilizing modern item detection models including Faster R-CNN, Mask R-CNN, YOLO NAS, YOLOv5s, YOLOv6s, and YOLOv8 variations. Comparative analysis evaluates detection accuracy, computational performance, and actual-time performance underneath difficult underwater conditions like low visibility and numerous particles kinds. A custom dataset simulates numerous underwater environments with photographs of particles of different sizes, textures, and occlusions. Metrics like imply Average Precision (mAP), precision, keep in mind, and inference time highlight YOLOv8x as the ideal model, attaining 92% precision and high recollect. The project includes a hardware layout for real-time deployment and an internet interface for comparative outcomes and interactive detection. This work advances dependable underwater particles control, promoting marine sustainability via scalable and green detection answers.