Infrared non-uniformity correction based on substrate temperature and Bayesian estimation

The advantages and disadvantages in nonuniformity correction method (NUC) based on calibration and scene of infrared focal plane array (IRFPA) were analyzed respectively. On this basis, a combined non-uniformity correction method was presented. According to the temperature of focal plane substrate at the moment of power on, previously stored gain and bias correction parameters were extracted from the FLASH for the corresponding temperature interval to eliminate the non-uniformity of detector preliminarily. After preliminary correction, a self-adaptive non-uniformity correction algorithm was presented in order to eliminate residual noises. The images after preliminary correction were decomposed by non-subsampled contourlet transform(NSCT), and the Bayesian threshold was used to estimate the signal and noise variance point by point. As a result, the residual non-uniformity noise was figured out and then got rid of. Experimental results show that such an algorithm could improve both the correction accuracy and the environmental adaptability effectively.