Matching radar and satellite images for ship trajectory estimation using the Hausdorff distance

A novel technique to obtain position, velocity and radar biases estimates of a ship is described by matching ship-borne radar images to geo-referenced satellite images. The matching is performed through the minimisation of the averaged partial Hausdorff distances between data points in each image. The minimisation rapidly yields robust geographical latitude and longitude position measurements, as well as ship heading and radar biases. The accuracy of the measurements is improved by feeding them into a Kalman filter that also allows estimates of the ship´s velocity to be obtained. The method provides an alternative effective position sensor for GPS denied environments, which may also be employed for automatic radar calibration of bearing and range biases or for indoor autonomous mobile robot navigation