Abstract: A new IRFPA nonuniformity correction algorithm is developed, based on the scene moving analysis. The correction procedure is to find the relationship between every two detectors using block matching of image sequence, and the matching information comes from the character of detectors' output signal along the time axis. The algorithm can not only adapt to the slow change of nonuniformity with time and with external conditions, but also eliminate the ghost that always apears in normal sence-based NUC analysis. The simulation results and the error analysis are also given.

Key words: IRFPA, nonuniformity correction, scene-moving-analysis, ghost