| Due to the limitation of manufacturing materials and technology, the infrared detector exist non uniformity. At present, the most common method is to use two points to correct the non uniformity. However, in order to maintain the good effect, the method needs to update the correction parameters by the calibration of shutter periodically.This will temporarily interrupt the detection process and there are a lot of limitations. This paper is to study the strategy of adaptive updating correction parameters, mainly focusing on the implementation of non uniformity correction of infrared image with no shutter calibration.Two points correction based on the assumption that the probe unit is time stable.Its correction parameters are fixed and can not adapt to the drift of detection unit response rate with the detector temperature, bias voltage, time change.so,It can not always maintain a good condition.In this paper, two kinds of strategies are derived to update correction parameters. The strategy based on detector temperature realize automatic updating of correction parameters by introducing the detector temperature and the base offset and using subsection Lagrange interpolation.The strategy based on the scene information updates the correction parameters by minimizing the square deviation of correction output with its mean filter.Based upon the above algorithm, a creative method of realizing the algorithm on FPGA is presented.To make use of respective advantages of FPGA logic and NIOS core,the correction model and data collection are realized by logic module and the calculation of complex correction parameters and system control are completed by NIOS programming.By parallel and pipeline execution of logic modules,and carefully designing bus architecture, the system can meet the requirements of high throughput.The contrast experiment with two point correction shows that the correction effect of the system is good in all scenes all the way.The system does not need the shutter, can continue to detect and volume is smaller and weight is lighter.At the same time, the system achieve a high throughput data processing by using single FPGA scheme.It is adapted to the developing trend of high frame rate and high resolution of infrared imaging technology.So,it is much of significance. |