Proceedings Article | 5 August 2009
Can-jun Xu, Lan Su, Gen-yuan Yang, Jin-song Zhao, Yi Cai, Shun-chen Pan
KEYWORDS: Image fusion, Polarization, Infrared imaging, Infrared radiation, Mid-IR, Image processing, Imaging systems, Polarimetry, Dielectric polarization, Polarizers
The initial images and the corresponding evaluation of a MWIR polarization imaging system are discussed in this
paper. This MWIR polarization imaging system consists of an objective, two folded mirrors, a polarimetric analyzer, a
relay lens, a MCT 320×256 FPA with 30μm pitch and 3.7~4.8μm spectral region. The effective focal length of the
re-imaging objective is 185mm, F-number is 4.0 and the polarimetric analyzer is located in the parallel light. When the
analyzer is rotated to four different degrees, that of 00, 450, 900 and 1350, a set of four original infrared intensity images is
acquired, then the Stokes vector's elements of I, Q, U and the degree of linear polarization P of the incident radiation
from remote scene can be solved. After that, several methods are taken to process the polarization information mentioned
above, such as the I-P fusion, the IQU fusion, the IQU pseudocolor fusion and the IQU pseudocolor fusion by color
adjusted. Respectively, the results of fused images are presented. The quantitative evaluations of the MWIR fused images
shows that the mean of gray scales has been improved 35%, the standard difference of gray scales has been improved
61% and the gradient of fused images has been improved up to 255%. The results of this paper indicate that the MWIR
fused polarization images are much clearer than that of the intensity images in profiles, edges and details of an object,
and it can obviously increase the contrast between object interested and the background. The different types of
polarimetric images can be displayed by means of image of fused images of different Stokes components gray scales or
images of pseudocolor fusion. These results offer multiple choices to the displaying of polarimetric images. The different
methods of polarimetric fused image can product different results. The results show polarimetric images contain new and
useful information that is not obtainable from the intensity images.
