Research On Dual Band Image Fusion Method

Fused through a specific algorithm, the same scene images of different mechanism sensors processed by dual-band image fusion obtain more comprehensive and reliable information than single-band image fusion. Due to its unique advantages,the technology were widely used in civil,military and other related fields. And how to improve the effect of the fusion image, increase the practicability of fusion method has been an urgent problem in the field of image processing. Some researches about the dual-band image fusion method were carried out in this thesis.On the basis of the existing fusion methods, the traditional laplacian fusion method was lack of fusion strategy selection and comprehensive description of image features. A laplacian fusion method based on structural similarity was proposed in this paper. Compared to the traditional laplacian fusion, a combination of image brightness, contrast and structural information was taken by the Laplacian algorithm based on structural similarity. Experimental results show that the fused images obtained by this method are more suitable for human eye system. The existing color image fusion method was lack of image contrast and prominence of targets. A color fusion algorithm based on target enhancement and contrast optimization was proposed in this paper. A separation colorization pattern was adopted in this method. First, the target was extracted from image by a segmentation algorithm based on significance, then enhanced by color coding function; color transfer based on the brightness and contrast optimization was applied for the background. Finally, making a combination for both the background and the target color fusion image, and it has been proved that the obtained color images were prominent and enhanced in experiments. In order to verify the practicability of fusion method, Combined with infrared dual-band fusion system, the affiliated PC software designed and compiled on visio studio 2010 platform could realize the serial control and image display processing. The effectiveness of fusion software and practicality of the system has been proved under the test of fusion method mentiond in the thesis.