The Optimize Technology Research On Image Fusion With Visible And Infrared Image

Image fusion is an important part of image processing system. Image fusion has been widelyused In the military, civil and other fields. But the current image fusion algorithm fail to take intoconsideration certain types of imaging mechanism and the specific image fusion of purpose. As aresult the fusion effect is not good. So this paper visiual image and infrared image fusion, andcarried out the the optimize technology research based on multi-resolution decomposition method.First of all, in the low frequency fusion stage considering that when the image scene iscomplex and there are multiple areas in the image need attention at the same time, The commonuse of low frequency fusion method and simple binary regional segmentation method can’teffectively reflect target to the fusion results, at the same time it will reduce the image contrast.This paper adopts segmentation method based on the information entropy of multi-threshold. Theimage will be segmented to multiple region which has similar temperature of a class of objectsinthe scene. And the next for each area fusion, so the result can effectively keep the clarity of theoriginal image and administrative levels feels.Secondly, in the high frequency fusion stage, according to the visual image and thecharacteristics of infrared image fusion, this paper chooses the the gradient rules which can mostlykeep of the Optical image rich details information. And according to the characteristics of NSCTdecomposition coefficient, this papet adopts the LOG which is scale variable to calculate thegradient. At the same time, the global gradient and consistency check have been worked. Themethod in keeping details information has better performance.Moreover, in view of the traditional image objective evaluation index existence andsubjective evaluation is not the same situation, this paper adopts the average gradient, spatialfrequency, structure similarity index and edge similarity indexs to comprehensive evaluationfusion results. And this paper combine the above indexs into a target by using the fuzzy integralmethod.The results show the objective evaluation index can effectively evaluate the image fusionresult case, and the subjective evaluation have good consistency.Finally, considering the traditional image fusion framework doesn’t consider multiresolutiondecomposition layers on fusion of the outcome, so its improvement, the improved framework canadaptive sure decomposition level of make optimal fusion results.This paper also builts up the image fusion processing software system, this system cancomplete source image storage and different method of image fusion, fusion results of storage andcompared, and other functions. The system has a good human-machine interface, user-friendly andfast execution speed.