Infrared Focal Plane Array Non-uniformity Correction Methods And Simulation

At first of this paper, the features and causations of non-uniformity in infrared imaging systems and sense of the scene-based NUC are described. Several basic methods of non-uniformity correction (NUC) for FPA, especially the two-point methods are studied and induced. Two-point compensation technique (TPNUC) are discusses and empirical are provided in the article. According the results of experiments, the two point techniques are compared. The advantages and disadvantages are evaluated. Following that, the combined correction algorithm is introduced, which has the advantage of the two-point NUC and ANN, the combined algorithm adjusts gain and offset correction coefficient automatically through neural network, so it can avoid the shift of two-point NUC, and adopt two-point calibrate correction coefficient as neural network iterative initial value, so it converges to the target value faster than conventional ANN correction algorithm. Based on the simulation results, the algorithm behave more satisfyingly. To avoid the disadvantage of the traditional spatial adaptive non-uniformity correction the low frequency dominated non-uniformity noise, a new way is used. At last of this paper, Simulation on CI system shows that the proposed scheme can achieve good performance when low frequency noise is dominant.