Publication

Abstract : The Search and Rescue (SAR) Marine operation is designed to help maritime workers identify the types of behaviors like water activities, ship monitoring, terrorism, malicious activity detection, etc. The marine environment is complex and changeable. The term "Maritime search and rescue" refers to actions taken to find, save any individual in the sea who gets lost, hurt, or killed outside, or as a result of a disaster caused by nature, technology, or people. The SAR team has performed various steps to identify drowning activity and rescue any individual suffering from distress. Drowning is considered to be one of the highest causes of death in water. Drowning leads to death, trauma, heart attack and seizures. It is very challenging and crucial to identify and save the person. Successful detection of drowning activity is essential when it comes to real time search and rescue operation. In this paper, deep learning model is developed to identify a drowning person. The network architecture used for creating the deep learning algorithm is Convolutional Neural Networks (CNN). Conv2dLSTM is used as a convolution layer in the CNN architecture that proves to have better efficiency for the proposed application. Deep learning model is successfully built using CNN architecture which has an accuracy of 0.9755. The processing time taken for the model to detect the drowning activity is 5s where the rescue operation can be done immediately to save the person.