On the Use of Visual Perception for Unmanned Surface Vehicles Navigation

We introduce the state-of-the-art on Visual Perception applied to the Unmanned Surface Vehicles (USV) navigation, pointing out the main challenges and directions in the topic. We use a systematic review approach including conference and journal papers and also thesis’ dissertations, with a defined set of keywords covering general robotic platforms (including USV) that attended to the visual navigation requisite. With a initial set of 59 works resulting from the initial search, we observed a number of works on the use of visual simultaneous localization and mapping (V-SLAM) for robotics in general. Nonetheless, a more detailed analysis showed that the use of V-SLAM and other visual perception techniques in autonomous surface vessels is a little researched topic, which provided 17 main contributions with only 4 strictly related to our particular topic, as described in the work. Our findings suggest that the use of visual perception can play an important role in the autonomous navigation of sailboats and other vessels operating on the water surface, and that it can still be more explored. Roboticists are developing such techniques that are making the application of computer vision and AI to robotics a very useful resource for enhancing the perceptual abilities of such USV platforms.