As developed in Chapter 3, FPA camera design trades blur with aliasing, both of which degrade the imager performance. Such analysis assumes a fixed spatial sampling rate at the camera focal plane, determined by the detector pitch of the FPA. If an FPA camera system can dynamically sample the optical image, then the FPA sampling rate no longer limits the resolution of the digitally sampled imagery.
There is considerable historical experience with dynamically sampled camera systems, particularly with first generation FLIRs. Simply scanning a 1-D column of detectors, or even a single detector element, can generate 2-D imagery with obviously higher spatial resolution than that of the single column or element. Such processing critically hinges on accurate knowledge of the scanning motion, which is often a precisely controlled function of the servo voltage that drives a scanning mirror assembly. Fusing such position information with detector intensity information generates the higher resolution imagery.
Two-dimensional micro-dither scanning can likewise increase the spatial resolution of an FPA camera. However such 2-D micro-dither scanning involves mechanical complexities beyond first generation scanning FLIRs. Such processing requires longer framing times, introduces jitter distortions, and increases the sensor weight and complexity: Precisely the problems that motivated the engineering of FPA cameras in lieu of scanning sensors!
As an alternative to controlled micro-dither scanning, the optical flow of an image sequence can be computed to tag a spatial coordinate with every pixel of the video sequence. Intensity and coordinate information can then be combined to create an image with higher sampling density than that of the FPA.
Online access to SPIE eBooks is limited to subscribing institutions.