Research On Nonuniformity Correction Algorithm For Infrared Image

In recent years,infrared imaging technology continues progressing,the core detectors continue updating,the detection systems continue becoming perfect day by day,especially the emergence of the infrared focal plane arrays with high sensitivity,high integration and low energy consumption,making the infrared imaging technology play an irreplaceable role in national security,environmental monitoring,national economic development and other major fields.However,the fringes and distortions in infrared images make the temperature sensitivity of the detector decrease and the detection distance shorten,which affect the image quality seriously.We call these fixed noise non-uniformity[NU]of infrared images.This paper aims to get the infrared images with high resolution,first of all,it describes the development and application of the infrared imaging technology,and the infrared detector,followed by the analysis of the causes,the presence forms and the general recognition of the non-uniformity noise in the infrared images.The mechanism of the non-uniformity correction process is pointed out.Finally,the paper studies the calibration algorithm and two typical scene-based algorithms.First,the statistical class,typical algorithms are:time-domain high-pass filtering,neural networks,statistical constant.Another is the registration class,typical algorithms are:motion compensation,algebraic method and inter-frame registration.The advantages and disadvantages of each are illustrated by comparison.In order to overcome the shortcomings of traditional calibration algorithms,such as blackbody reference and strict linearity,the paper proposes a nonuniformity correction algorithm based on the focal plane mean response.The experimental results show that the new algorithm has better linear performance and the second-order function is better than the second-order fitting in the second-order function image,and has a strong non-linearity.In order to overcome the shortcomings of traditional registration algorithm,such as the difficulty of registration,the large amount of computation and the difficulty of hardware implementation,the paper proposes a recursive matrix correction method.The experimental results show that the new algorithm has the advantages of small computational complexity,low requirement for storage space,adaptive adjustment of parameters,good convergence and stability,and high image visibility after correction.