The Role of Vision for Underwater Vehicles

{ Diierent sensing techniques have been used for a long time in the development of remote operated vehicles (ROVs) and, in the recent years, autonomous underwater vehicles, (AUVs). Sonar has been the most popular choice in what concerns depth information or obstacle detection in most of the existing systems. Vision, however is a powerful sensing modality and could be used in many tasks where accurate measurements at a short range, are needed. This paper discusses the possible use of vision in the context of autonomous underwater vehicles. Tasks, such as object avoidance or recognition , grasping, docking, sea bed reconstruction, underwater surveillance, inspection, cable maintenance , are among the set of those where computer vision may have an important role. As an example, we present a vision system designed for recursive depth estimation based on a underwater vehicle equipped with a camera. Results obtained with underwater images are presented. The technology of autonomous or semi-autonomous underwater vehicles has come to a point where very performing systems have been developed. Sonar has been the most widely used sensing modality for tasks such as depth extraction or obstacle detection. However, vision is a sensing modality which provides very high resolution, yields accurate depth measurements specially at short range and, therefore, could be used for a set of tasks in autonomous underwater systems. On the other hand, Computer Vision has now grown to a mature point where robust, sophisticated systems have been developed and are able to operate at fast speeds. Besides depth recovery, which could be a goal on itself, interesting tasks that could be addressed by using vision techniques are surveillance, recognition, visual-guided docking or manipulation, obstacle detection 5], trajectography, low altitude navigation 11], and so on. For docking or manipulation, vision can directly provide feedback information to drive the motor actions of the underwater vehicle. The coordination of vision and action is, in fact, one of the core ideas of the recent research in the so called active vision 1]. Bathymetric information can be used for diierent purposes but in a similar way, one could use vision to estimate the vehicle position in a known environment. In the last few years, computer vision researchers have designed diierent obstacle avoidance techniques which tend to be robust, fast and are suitable for underwater application. Naturally that the underwater environment ooers some new challenges when compared to land robotic applications. For this reason, …