Deep Neural Network-Based Visual Identification of Naval Surface Vessels
Sean McCormick, Violet Mwaffo, Donald H. Costello · 2024
The United States Navy plans to expand its inventory of uncrewed aerial systems (UASs) and expects these systems will have the ability to operate in maritime environments where radio communications will be restricted. To operate in these environments the UASs will need to have the capability to navigate autonomously, without a human in or on the loop to control the UAS. Some form of visual navigation seems to be a plausible solution. This would require a UAS to visually differentiate between different ships. This paper investigates the possibility of a deep neural network model to classify and differentiate between similar surface vessels in the maritime environment.