Research On Influencing Factors And Fusion Algorithm Of Infrared Polarization Imaging

The infrared detecting technology is one of the most important methods in target detection, tracing and recognition. But the performance of infrared detecting would go down sharply if there were little temperature differences between the targets and the background or the targets got camouflage. The targets’polarization state was decided by the speciality of the target itself. As a new kind of imaging technology, the polarization imaging technology use targets’polarization state information to get a contrast image, and it had a better performance for detecting camouflage targets. The infrared polarization imaging combined the infrared with polarization technology, which could hugely improve performance of target detecting and recognition systems in complex working conditions. This paper studied the principle, the influencing factors and the fusion algorithm of the infrared polarization technology, wishing to solve the problem of infrared system which may not find the true targets under complex working conditions.This paper studied the polarization feature of the targets’infrared spectrum, and finded the model of targets’infrared polarization feature. This paper also simulated the main factors which would effect the performance of the infrared polarization imaging. A new image fusion method based on image evaluation was given out, which could improve the targets’contrast and image intensity by50%-400%and10%-100%compared to HSV fusion and YCbCr fusion.The main research content and achievement include:1. The principle of reflex polarization production was studied by Fresnel Reflection Law; the infrared polarization model was deduced by using Kerchief’s law and Planck blackbody formula; also, the Atmospheric depolarization of infrared polarization and the theory of polarization detection were studied.2. According to the infrared polarization model, several factors that may affect the infrared polarization imaging were deduced and simulated, such as the incidence angle, the refractive index, the wavelength, the light regimes and the roughness value of the target. The confirmatory experiments were done to find the conditions for the system.3. The image fusion algorithm was studied. Several image preprocessing methods were studied which were used for image normalization and anti-noising. The ideas of3fusion algorithms based on color-space were researched and contrasted. Several image evaluating criterions were introduced. The paper gave out a new kind of image fusion algorithm based on image evaluation.4. A experimental system for infrared polarization imaging was set up and several outdoor experiments of different targets with different distance were taken. The result showed the fusion algorithm based on image evaluation could enhance the contrast and the image intensity effectively.