NSST And Adaptive Blocking Based Multi-focus Image Fusion

With the advent of various image sensors with different functions,the limitations of these sensors in terms of space and imaging conditions have also emerged.In response to this problem,a kind of image information obtained by multiple sensors has been integrated into one pair.The fusion technology of the image is not subject to the constraints of time and other conditions.This method not only solves the limitation of sensor imaging,but also integrates the information to improve the image accuracy.Due to its good application effect,in recent years,in order to obtain a better fusion effect,a lot of researches have been carried out.The two fusion algorithms proposed in this paper are based on this purpose.The main content and innovation of this article:(1)The background,existence and research status of image fusion are elaborated.The classification of image fusion is introduced from two levels of information level and image source respectively,and then the most commonly used pixel level among the three fusion levels.The fusion method is introduced from the two aspects of transform domain and airspace,including common pyramid transform,wavelet transform and Contourlet transform.Finally,the existing problems are proposed.Based on these problems,two fusion algorithms are proposed.(2)The first algorithm is a fusion algorithm under the framework of NSST and DWT.This method combines the advantages of the two methods,which not only retains the excellent ability of NSST in edge detail processing,but also retains the advantages of DWT in conformity with human visual characteristics.Moreover,this algorithm also has good translation invariance and directionality,which solves the shortcomings of DWT and preserves the good performance of NSST itself.The whole algorithm first performs NSST decomposition on the source image and then performs the decomposition of the decomposed low-frequency components.DWT transform,then use ISNL and adaptive block combination and local wavelet energy as the low-frequency and high-frequency fusion rules to complete the fusion.Comparative experiments show that the effect of this algorithm is significantly better than NSST and DWT(3)The second algorithm is proposed based on the problem of high frequency and low frequency isolated fusion on the basis of the first algorithm.The fusion of high frequency and low frequency is solved to solve this problem.The fusion of the low-frequency part is fused by the adaptive block method.The adaptive block-based algorithm uses the differential evolution algorithm.Due to the defects of the method itself,the source label map is used to accurately process the fusion result to obtain the low-frequency fusion.Coefficient;the fusion of the high-frequency part uses the local average gradient algorithm.When the high-frequency coefficients are fused,the corresponding low-frequency coefficients are also obtained,and then the obtained coefficients of the low-frequency coefficients and the low-frequency coefficients are weighted averaged.The low frequency fusion coefficient is obtained.The above is the improvement of the algorithm on the fusion rule.This improvement solves the problem of isolated fusion.Through comparison experiments,it is found that the results obtained by the algorithm are improved both subjectively and objectively.