Deep Learning Based Automatic Plastic Recognition Using Autonomous Underwater Vehicle
Plastic pollution in water bodies is one of the biggest issues in this era. It affects the marine animals in turn affecting the whole biological chain. The plastic suspended in water is consumed by many organisms which leads to their death which causes the biological imbalance. It also affects humans as drinking water can also be contaminated. So, the main motive of this project is to reduce plastic waste in water reservoirs. This paper discusses number of algorithms to detect underwater debris efficiently. Algorithms include YOLO, tiny YOLO, Faster RCNN, forward-looking imaging sonar, etc. The dataset used for training taken from JAMSTEC E-library of Deep-sea Images. This paper also discusses about the issues in underwater video processing. We have designed an underwater autonomous vehicle for detection of plastic, design of which is also discussed in this paper. Thus a complete system to detect plastic is discussed in this paper.