Image Filtering And Fusion Based On Lifting Wavelet Transform

The image will inevitably be contaminated by noise during collection or transmission.The presence of noise in the image not only destroys the uniformity of the image area,but also is not conducive to the preservation of the image edge information;Image fusion is a process in which the same target is reconstructed by the same sensor obtained from the same sensor at different angles or different sensors,and the redundant part is removed to obtain a complete and complete target object.Image fusion has a wide range of applications,such as military,medical,and intelligent identification.According to its fusion method,it can be divided into pixel level,decision level,and feature level.Pixel level image fusion is favored by scholars and experts because of its simple operation and intuitive method.Decision-level,feature-level image fusion is mainly applied in the transform domain.In general,image filtering is the preprocessing of images.Image noise seriously affects the subsequent processing of images,such as image fusion,image segmentation,image enhancement,image recognition,and so on.The effect of image filtering will affect the effect of subsequent image processing to a certain extent.In this paper,lifting wavelet is used as a transformation tool for image filtering and fusion processing.The advantages of lifting wavelet transform mainly include in situ computing(In-place),invertibility,simple computation.It also inherits all the advantages of wavelet.The object of filtering in this paper is ultrasound image.The noise of ultrasound image is mainly dominated by multiplicative noise.Experiments show that the lifting wavelet and threshold filter have better denoising effect on multiplicative noise.This article uses the filter evaluation method and proposes a filter threshold function.In image fusion,image fusion objects are not only limited to ultrasound images,but also have good effects on optical images.This article proposes fusion methods and evaluation criteria.Through the above innovative methods of image filtering and fusion,the experimental results show that compared to the commonly used methods,the algorithm has a better filtering and fusion effects.The main content of this article is as follows:(1)In the ultrasound image filtering process,the high frequency and low frequency parts are obtained using a lifting wavelet transform.The noise is generally concentrated in the high-frequency wavelet coefficients,so as long as the high-frequency coefficients are considered,the low-frequency coefficients remain unchanged.The filtering method summarizes and improves the advantages of various threshold functions,and is beneficial to the threshold function of ultrasound images.The filter noise is evaluated and compared with other methods.The innovative filtering method of this paper is better for the edge retention of the target area,and the dark area and the reverberation area are relatively smooth.(2)In image fusion,this paper uses the lifting wavelet transform and S operator to fuse images.Lifting wavelet transform results in low frequency wavelet coefficients and high frequency wavelet coefficients.The wavelet coefficients are weighted,the high-frequency coefficients are processed by S algorithm,and then the wavelet coefficients with relatively large signal components are selected for fusion.This article first mentioned the S algorithm and discussed the S algorithm in detail.The S algorithm can effectively evaluate noise and signal components.Comparison with other methods can show that the proposed fusion method is better in both subjective and objective evaluation.This paper proposes that the image filtering method is mainly applicable to ultrasound images,and it needs further demonstration for other images.In terms of image fusion,the new algorithm is applied not only to ultrasound images but also to optical images.However,there are strict requirements on the fused image,and the original image must be pixel-by-pixel.