Computational Color Imaging Based On The Color Look-up Table Of Multidimensional Information

In recent years, the continuous development of shimmer night vision and infrared thermography in the military and civilian areas have been widely used, their imaging mechanism is different, so can provide the complementary information of the scene and objectives.Therefore, the computational color imaging based on the shimmer/infrared natural sense can enhance the perception of the scene and improve the target detection and recognition ability. This paper introduces the relevant theory and historical development of the computational color imaging, and launches a research about the color look-up table algorithm. The color look-up table has a better sense of the natural effect, intuitive color performance and improves real-time characteristics, etc. In this paper, different sizes, different regions and different resolutions multidimensional information are presented in the color lookup table, which means that the local information, texture information, and pixel correlation are applied to color remapping. The results show that natural sense images are got with more rich details and the smaller distortion by the improved method.The major works include:(1) The improved method for optimizing two dimensional color look-up tablesCommonly used optimization of the two dimensional color look-up table is european filling method. This article focuses on the distribution of the color pixels, takes a different starting point for different filling look-up tables, proposes the limited pixel neighborhood mean filling method to better populate the look-up tables, increase the dynamic range of the look-up tables by processing the image previously; proposes the luminance permutation method based on YUV space, that effectively solved the problem of the misinformation.(2) The dual-band natural sense fusion method based on Laplace pyramidsBy studying the basic principle of the Laplace pyramid decomposition and its application in a variety of image fusion. By applying the theory of Laplace pyramid to the color look up table, the image with more spatial resolution are got.(3) The dual-band natural color fusion method based on local mean and varianceThis section analyzes the color transfer method with the pixel neighborhood grayscale mean and variance as the transfer parameters from color image to gray image. The local mean mainly reflects the luminance of the image, and the variance reflects a certain texture information. Using certain texture information to pixels matching, color transfer, and accurately control of the color transfer way between two images, a good color fusion effect is obtained.(4) The algorithm of multi-dimensional information color look-up table based on image segmentationDifferent transfer factors having different advantages and disadvantages, so the reference image is divided by the K-means clustering algorithm. According to different divided parts, the details can be strengthened and the distortion can be reduced by using different color look-up table.