A Novel Approach for Non Uniformity Correction in IR Focal Plane Arrays

Abstract

High exigency of sophisticated state-of-the-art Infra Red (IR) cameras has witnessed a proliferation of larger format IR Focal Plane Arrays (FPAs) arranged in a 2D grid of photo detectors placed at focal plane of imaging system. Current IR FPAs are performance-restricted by variations in responses of individual detector elements resulting in spatial non-uniformity even for a uniform radiation source as scene. Generally, two-point Non Uniformity Correction (NUC) is the preferred technique to cater for this problem. This technique is limited by variations in performance over the entire operating range causing the residual non uniformity (RNU) to vary more on moving farther from the calibrated points. In the present approach, an integration time based NUC method is proposed using least square regression; wherein photo responses of individual detector elements at different integration times are measured when the IR camera is subjected to any uniform radiation source. A close linear approximation of these responses at various integration times is examined in the form of the best fit line by minimizing sum of square of errors. Consequently, a gain and offset value for each detector element is generated and stored in memory, which is utilized during real time to display the NUC corrected image. The results of present INT-LSR-NUC exhibit a considerable gain in performance when evaluated against conventional two-point NUC.