Acquisition And Fusion Of LWIR Polarized Images

Infrared imaging technology is a powerful detection method,but it has obvious disadvantages: the edges are blurred;details are not prominent;intensity distribution is centralized and so on.In the complex background and infrared stealth condition,the infrared imaging method cannot detect objective effectively.The object itself determines the object’s infrared polarized character.Infrared polarization system can make up infrared imaging deficiency by obtain polarized information.It has advantages in detecting objects with small temperature difference,as well as infrared stealth objects.We had a study on character of infrared polarization image and image fusion technology.The main work is as follows.(1)Analysis of infrared polarization image characters.The infrared radiation polarization model was established based on infrared and polarization principles.The effect of incidence angle,refractive index,roughness and wavelength were analyzed.This established essential theoretical basis for infrared polarization object detection.(2)Establishment of time division infrared polarization system.Units of infrared polarization imaging,stepper motor control and infrared polarization images acquisition were linked together,debugged and packaged.System architecture is simple and convenient to carry out.(3)An improved multi-scaled image fusion method based on support vector transform(SVT)was proposed.The low frequency coefficients which were fused with weight averaging method while the band pass coefficients were fused based on local energy matching method.Because distributions of source images’ support vector coefficients have a large difference,the matching degree is introduced.(4)Infrared polarized image acquisition and fusion experiment.Infrared polarized images of different scenes were obtained by the established system,and the infrared polarization images were calculated using the Stokes vector.Then the characters of infrared polarization image were concluded.The outlines and details of infrared polarization images are obvious,and the polarization between different material objects can be shown in images.However the infrared polarization images have the disadvantage of low contrast.At the end,effects of different image fusion algorithms were contrasted,and the evaluation proved that the proposed methods were more effective.