Object Tracking for an Autonomous Unmanned Surface Vehicle

The conventional algorithm used for target recognition and tracking suffers from the uncertainties of environment, robot/sensors object, such as variations in illumination viewpoint, occlusion seasonal change, etc. This paper proposes a deep-learning based surveillance reconnaissance system unmanned surface vehicles by adopting Siamese network main neural architecture to achieve tracking. It aims detect track suspicious targets. proposed perceives surrounding environment avoids obstacles while is evaluated with accuracy, precision, recall, P-R curve, F1 score. empirical results showed robust vehicles. approach contributes intelligent management control required today’s ships, also provides new