| Polarization imaging has become one of the frontier directions of contemporary imaging technology development,which is mainly used in target detection and other military and civilian applications.Because color can improve the contrast between the target and the background,so that the target can be better detected,the development of color polarization in target detection is also more and more rapid.In order to make the scene of polarization imaging have natural color after processing,and to highlight the detection target better,a color polarization display technology based on color transfer is studied.The theory of polarization stokes vector parameter and the theory of polarization degree of reflected light are analyzed.The experimental system of polarization imaging is studied,and the experimental device of color polarization image acquisition is built by combining the existing experimental equipment.The multi-field and multi-scene multi-group color polarization image acquisition experiment is carried out,which makes full preparation for the algorithm verification of the subsequent research.In this paper,the polarization characteristic image is mapped to HSV space by using the algorithm.Combining with the selected reference image which is the same as the scene color,the standard deviation and mean are used to correct the image color in YIQ color space,and then the image is converted back to RGB color space.Two kinds of methods,no reference image and full reference image,are used to evaluate the image after color transfer and to calculate the color transfer with the predecessors.The method was compared.The results show that,compared with the previous algorithms,the proposed algorithm improves the global image contrast by 33.68%,the standard deviation by 31.19%,the information entropy by 14.57%,the average gradient by 24.46%,the peak signal-to-noise ratio by 14.78%,and the structural similarity by 28.5%.3%.The validity of the algorithm is verified by the quality evaluation data. |
PDF Full Text Request