Prioria Robotics (United States)

Knowledge of the distance between a ship and other objects on the horizon has important implications for safety on the open ocean. Determining this distance can be challenging. RADAR sensors, which represent the standard solution to this problem, require a great deal of cost and power. A structure-from-motion computer vision technique is presented here to integrate known camera motion and a single pixel track on an imaged object to estimate range to that object. This approach leverages inherent wave motion of the camera platform along with a global exponentially stable observer to incrementally converge upon a range estimate. This method is detailed herein, and presented with both simulated and real-world results. This provides a passive and inexpensive solution to the open ocean ranging problem.