Underwater Marine Life and Plastic Waste Detection Using Deep Learning and Raspberry Pi

Abstract

Computer vision techniques have become increasingly popular in recent times. One of the applications of computer vision is the detection of objects in an image. In an object detection system, a target object is detected, localized, and labeled on the basis of the class of object. Detection of objects in harsh environments, accuracy in the detected objects, and speed at which they are detected are some of the major challenges in computer vision technology. Underwater imaging has its own set of difficulties namely: reduced visibility; objects possessing highly deformed shapes (plastic bags); means of communication of the captured image for proper processing. In this work, development of a system to detect marine life and plastic wastes in underwater environments is proposed. Deep learning technique is applied on images for the detection and localization of the desired target objects. A table comprises of the parameters of machine learning such as accuracy, precision, recall, F1 score, misclassification rate, true positive rate, false positive rate, true negative rate, and false negative rate.