Study On Double-band Color Image Fusion Algorithm Based On2-D Color Look-up Table

With the development of photoelectric imaging technology and modernization of themilitary equipment, the multiband imaging is increasingly applied to technology research andequipment field, especially the fusion of double-band image color in infrared and visible light,has been rapidly developed over this years. In recent times, the color fusion technology hasbeen developed from the process of false color assignment to natural sense color transfer, andgradually from color transfer as a main method to the color image fusion, and greatachievements has been obtained. Although domestic research in this field started later, it hasbeen developed rapidly in recent years, especially in Beijing Institute of Technology (BIT), alot of outstanding contributions have been made.This paper has summarized the previous work, and established the color transfer as thefoundation for the research, and we have further studied the color transfer theory and relatedcolor spaces. In this paper we have studied the image fusion algorithm based on2D colorlookup-table with TNO method by Holland and BIT method by China as the representative,and improved the algorithms on the aspects as little correlation between source image andlookup-table, lack of fusion image color, and obviously influence of reference images. Wehave improved the color image fusion algorithm based on2D color lookup-table, focused onthe procedure of the algorithm, and presented a multi group of comparisons under differentscenarios with TNO method and BIT method to validate the natural fusion effect of thisalgorithm.In order to improve the situation of great influence and manual choosing of referenceimages, we have put forward a fusion algorithm with several reference images, to make thecolor of a fusion image originated from difference color lookup-tables, and to increase thestability of natural sense of color. At the same time, in order to facilitate the automaticmatching of multi color lookup-tables, we also have proposed the partial assignment methodof block colorize. Finally, we have analyzed different images by changing several influencefactors, to verify the effectiveness of the algorithm.The algorithm proposed in this paper has improved the shortcomings in the existingmethods by richer color and better natural feeling, provided a technical support to thedevelopment of color night vision technology.