Infrared And Visible Light Image Registration And Fusion Technology Research

With the rapid development of imaging sensor technology and more and more complex actual service environment, a single imaging sensor is difficult in meeting the current requirements of use, while multiple imaging sensor system, which gains complementary or redundant information from every imaging sensor can meet the practical needs. Many multiple imaging sensor systems have been developed. The system which combines infrared and visible image was mainly discussed in this thesis and the registration technology、fusion technology of infrared image and visible image were studied emphatically.The main contributions of this thesis are:(1) Based on the basic theory and method of image registration, according to the characteristics of infrared and visible image, a registration method for infrared and visible image was proposed which combined marginal alignment and mutual information in this dissertation. Firstly, the edge image of the infrared and visible image were detected based on the maximum module of wavelet transform. Then the alignment metric and normalized mutual information were calculated respectively. And the new similarity measure function could be obtained by the self-adoptive weighted average of Alignment Metric and normalized mutual information. Finally, transform parameters of the images to registration were obtained by calculating the extreme of the proposed similarity measure function. Experiment results showed that this method could reduce the time needed for registration, and had higher accuracy and robustness.(2)By studying the basic theory of image fusion and multi-scale geometric analysis and intelligent image fusion method, the Non-Sub sampled Contourlet Transform (NSCT) was introduced to infrared and visible image fusion. Firstly, the images were transformed into low frequency sub-band and several high frequency sub-bands. Then the low frequency sub-band adopted the fusion rule of regional weighted average energy method, while the fusion rule based on Pulse-Coupled Networks (PCNN) was used for high frequency sub-bands. Finally, the fusion image was obtained by the inverse NSCT. The experiment results showed that the fusion algorithm proposed in this thesis could achieve better effect than other fusion algorithms.The obtained results provided a new technical approach for the practical application of infrared and visible image fusion by which the good image fusion method could be developed.