Image Fusion Algorithm Based On Wavelet Transform

Multi-focus image fusion is an emerging technology in the field of image fusion, it can be used to solve the image blurring problem due to the focus is not on the whole object, it further facilitates the image processing analysis. Wavelet transform is a rapidly emerging technology. It can be used to transform the image signal into the frequency domain to do more in-depth study by taking the characteristics of multi-resolution analysis.This paper mainly studies the wavelet-based multi-focus image fusion algorithm. Multi-focus image can be decomposed into the approximation and the detail information of the image based on the wavelet transform theory. The human visual system is highly sensitive to the changes on the contrast of the local detail information. In order to obtain a better fusion result, the wavelet transform and the local contrast of the image are combined to solve the proposed problem. The main contents are as follows:(1) The wavelet transform theory and multi-focus image fusion algorithm have been analyzed. An algorithm for multi-focus image fusion has been developed based on the frequency domain analysis, i.e. perform the wavelet transform to the image and select the proper fusion rules in the frequency domain;(2) An adaptive multi-focus image fusion algorithm has been analyzed. This algorithm overcomes the disadvantages of a simple maximum value selection algorithm. The factor which is relative to the wavelet coefficients is defined as the weight coefficient for calculation in high frequency range. Finally, a fusion image which contains the useful information of the source image can be obtained;(3) A multi-focus image fusion algorithm based on the contrast has been analyzed. This algorithm is based on the local image contrast theory. The contrast of the high frequency wavelet is defined. Certain rules have been adopted to calculate the wavelet direction contrast in order to obtain the fusion image which can meet the requirements of the human vision characteristics;(4) An improved adaptive multi-focus image fusion algorithm based on the contrast of the wavelet direction has been proposed. This new algorithm takes the advantages of the adaptive multi-focus algorithm and the contrast-based multi-focus image fusion algorithm. It can retain the useful information of the source image which participates in the integration; and meet the characteristics of the human visual system. The effectiveness and feasibility of this new fusion algorithm has been proved;(5) The optimum value of wavelet decomposition layers has been discussed. By comparing the results of various levels of fusion and the experimental data, the optimum number of wavelet decomposition layers has been determined.