Research On Non-uniformity Correction Algorithm For Infrared Imaging System

With the improvement of the device structure design, the manufacturing technique of theInfrared Focal Plane Array (IRFPA) is growing mature. Nowadays IRFPA is developing fasttowards to realize large scale of pixels, high detection sensitivity and small size of pixels.This tendency leads infrared products wildly used in extensive fields. While thenon-uniformity noise in steering focal plane array imaging system badly affects the qualityof image, so it is important to find appropriate non-uniformity correction (NUC) algorithm.NUC algorithm can be roughly classified to two kinds: one is based on calibration method,the other is scenario-based method. Both approaches have their own advantages anddisadvantages. In this thesis, a cooled infrared camera was used to image realtime andacquire data. Algorithm based on calibration and temporal high-pass filter conductedin-depth research. The main contributions are given as follows:According to imaging process of infrared system, this thesis sets up the response modelof each pixel in IRFPA, analyzes the relation between incident light radiation intensity andthe output totally, also finds relation between the temperature of blackbody and radiationintensity, then do the research on the self-radiation of the optical system.To fitting the curve which represented the relation of different factors and thenon-uniformity it inducted, lots of experiment did in the thesis.This thesis brings forward two kinds of calibration method to realize the non-uniformitynoise which conducted by the ambient temperature. One is using offsets obtained bydifferent ambient temperature. The other one is get the correction coefficient by adaptionmethod.The temporal high-pass filter correction method also researched in this thesis. Algorithmdeduced by theories of scene comlexity analysis in image and spatial low-pass filter, do thedigital image data selection correction to elimiate the ghost occurded in original algorithm.