Multi-focus Image Fusion Algorithm Based On NSST Transform

In recent years,image fusion has become an important topic in the field of image processing.Image fusion technology is widely used in computer vision,monitoring,medical imaging,remote monitoring and other fields.Among them,multi-focus image fusion is an important branch of image fusion,and has been a research hotspot.For multi-focus image fusion,a lot of research and algorithms have been proposed.With the advent of big data era,conventional wavelet analysis can not provide flexible multi-resolution and multi-directional decomposition to meet the needs of large amount of image fusion.Nonsampled Shearlet Transform will not be limited,and has translation invariance and strong directional selectivity.At the same time,it has better image processing performance,which can better meet the needs of subsequent processing.This paper is based on the special performance of NSST transform,and deeply explores(1)The second chapter introduces the research significance and development status of NSST based image fusion,and analyzes the basic theory of image fusion,related theoretical framework,a variety of commonly used transformation algorithms and different types of evaluation index.In Chapter 3,the two steps of NSST transform are deeply studied: Non-Subsampled Pyramid Filters and Shearlet Filte Banks,and the image fusion framework of NSST transform.(2)In Chapter 4,an image fusion algorithm based on NSST transform and Spiking cortical model is proposed.In the first step,the source image is decomposed by NSST;in the second step,the fusion rules are selected according to the conditions;in the third step,the image is reconstructed.A large number of experiments and analysis show that the method in this chapter can effectively solve the problems of image distortion and artificial texture caused by traditional algorithms in multi-focus image fusion.(3)In Chapter 5,an image fusion algorithm based on NSST transform and improved sparse representation is proposed.Firstly,NSST transform is applied to the source image to obtain the low frequency and high frequency coefficient matrix.Then,for the low frequency coefficient matrix,the fusion rule of low frequency coefficient based on SR is used;then,for the high frequency coefficient,the fusion rule of integrable energy sum is proposed.Finally,the fusion image is obtained by inverse transform.Experimental results show that the algorithm can retain more detailed information and has certain advantages in visual quality and objective evaluation.