The Fusion And Evaluation Of Multi-Focus Images Based On Wavelet Transform

Wavelet transform is an effective means to process images processing .It has been used widely in a lot of fields of image processing .For instance the image is strengthened, compressed, analyzed, because it is MRA (multi -resolving-analyze) in time and. frequency fields. Today, not only wavelet theories go further more perfectly, but also closer to applications.This paper tries to merge wavelet theories with the multi-focus image perfectly and get better fusion result through studying and practicing of the basic theories of image processing, especially wavelet theories and the fusion of multi-focus images. This paper does researches of the fusion of multi-focus images based on wavelet transform mainly in two ways, explains it as follows:Firstly, base on exiting study, such as algorithm evaluating wavelet -image-fusion and traditional non-wavelet image fusion algorithms, and with my own practical experience on evaluating multi-focus image fusion, this paper improves the existing evaluating algorithm - wavelet energy algorithm, makes the wavelet function can exert its advantage which is MRA fully,Then, to analyze and verify their functions and effects, many kinds of wavelets have been selected in this paper as mother-wavelet functions for image fusion, such as Haar, Db10 Sym8, Coif5, Bior6.8 and Dmey wavelet .The strategy of image fusion is based on pixel to pixel. A set of evaluation method such as information quantum, characteristics in statistics and SNR have been adopted, the wavelet energy algorithm is especially presented for evaluating image fusion. The advantages and disadvantages of wavelets in image fusion have been obtained through the subjective comparison of the fusing pictures and objective evaluation of experimental data .The result can be used as a guide for wavelets selection.This paper evaluates the improved wavelet fusion algorithm by means of several kinds of evaluating algorithms from several points of view. Finally, it is concluded the improved wavelet fusion algorithm has made better result ofimage fusion, whether from the objective data that are tested, or from subjective observation.