Stereo spatial super-resolution technique for multiple reentry vehicles

We present a new technique to spatially super-resolve closely spaced objects (CSO) consisting in this case of multiple reentry vehicles released during the midcourse stage of an ICBM. Closely spaced in the context of this paper refers to objects that look like a single blob in the focal plane of the corresponding sensor, with separations of a pixel or less. The technique presented here uses multiple time-consecutive frames from two or more sensors staring at the same CSO complex and exploits the time intensity variations among the reentry objects and the geometric diversity of the observers. The algorithm utilizes time resolution to obtain improved spatial resolution. Although the super-resolution technique can be applied individually to each focal plane, the line of sight (LOS) intersections needed to locate the objects in 3D introduces undesired ghosts. By combining both focal planes via stereo processing, the new approach eliminates the ghosts and provides super-resolution in the 3D volume selected for analysis. We have tested this new technique in a high-fidelity CSO/focal plane simulator with good results.