Development of Autonomous Underwater Vehicle Equipped with Object Recognition and Tracking System

摘要

The development and design of autonomous underwater vehicle (AUVs) provides unmanned, self-propelled vehicles that are typically deployed from a surface vessel, and can operate independently for periods of a few hours to several days. This project discusses the development of an AUV equipped with object recognition and tracking system. In this project, the motion of AUV is controlled by two thrusters for horizontal motions and two thrusters for vertical motions. A Pixy CMUcam5 is used as a vision sensor for the AUV that is utilized to recognize an object through its specific color signatures. The camera recognizes an object through colour-based filtering algorithm by calculating the colour (hue) and saturations of each red, green and blue (RGB) pixel derived from built-in image sensor. When the camera recognizes an object, the AUV will automatically track the object without any operator. Preliminary underwater experiments have been carried out to test its ability to stay submerge underwater as well as its functionality to navigate and recognize object underwater. Experiments also have been carried out to verify the effectiveness of Pixy CMUcam5 to recognize a single and multiple objects underwater, then tracks the recognize object. This work reports the findings that demonstrate the usefulness of PixyCMUcam5 in the development of the AUV.