Focus3D: A Practical Method to Adaptively Focus ISAR Data and Provide 3-D Information for Automatic Target Recognition

To improve ATR identification of ships at sea requires an advanced ISAR processor - one that not only provides focused images but can also determine the pose of the ship. This tells us whether the image shows a profile (vertical plane) view, a plan (horizontal plane) view or some view in between. If the processor can provide this information, then the ATR processor can try to match the images with known vertical or horizontal features of ships and, in conjunction with estimated ship length, narrow the set of possible identifications. This paper extends the work of Melendez and Bennett [M-B, Ref. 1] by combining a focus algorithm with a method that models the angles of the ship relative to the radar. In M-B the algorithm was limited to a single angle and the plane of rotation was not determined. This assumption may be fine for a short time image where there is limited data available to determine the pose. However, the present paper models the ship rotation with two angles - aspect angle, representing rotation in the horizontal plane, and tilt angle, representing variations in the effective grazing angle to the ship.