Underwater Object Detection Using Image Processing

Abstract

This project describes the architecture and design of an improved configurable underwater object detection using image processing. The vision based method for detecting the objects under water is carried out using Yolo based convolutional neural network (CNN). YOLO (You Only Look Once) is a real time object detection algorithm. This algorithm applies a single deep neural network to the full image, and then divides the image into regions and predicts bounding boxes and probabilities for each region. These bounding boxes are weighted by the predicted probabilities. The main contribution of this project is to detect the objects and send the captured image and its real time coordinates to the specified mail address.