Watercraft-Net: A Deep Inference Vision Approach of Watercraft Detection for Maritime Surveillance System Using Optical Aerial Images
Alvin Sarraga Alon, Jonel R. Macalisang, Ryan C. Reyes, Rovenson V. Sevilla, Gemma D. Belga
Abstract
The ocean and any form of bodies of water must be protected and secure from any intruders and to monitor our ocean, technology must be used and integrated for more efficient monitoring. Automatic boat detection plays an important role in maritime surveillance. However, the maritime environment represents lots of challenges such as the wave of water, boat movements, and weather condition. This paper presents a method for detecting moving boats from a sequence of images using a deep learning approach. In this study, the researchers proposed a detection system for the boats in the ocean using optical aerial images. The researchers conducted testing and the results were favorable. Upon testing the researchers obtained a 90% accuracy of detection of the ship in the ocean using the single images, video feeds, and live feeds. The experiments show promising results.