Improvements On Multi-Source Image Fusion Algorithm Based On Transform Domain

Using information technologies and methods to help social economic organizations to improve the efficiency and level of modern management is one of the hot topics in the field of management science and engineering.Image fusion is of important application value in management,which can improve the efficiency of management and help to manage more scientifically.If image fusion is applied to monitoring,the effectiveness of monitoring can be improved.Image fusion can also be applied to express security inspection which can do effectively identification of prohibited items,timely checking,ensuring safety and improving the accuracy and efficiency in express security inspection.With the growth of computer processing speed,the data stored in the form of video is playing a more and more important role.To get information from different aspects of the same scene,different types of source images can be acquired by different sensors.So,the research and application of multi-source image fusion is very necessary.There are some problems to be solved in multi-source image fusion,such as the obvious block effect,edge blurring or missing information in the fusion image.How to overcome these problems to obtain more high performance image fusion results needs to be solved.This paper has done some researches and explores of multi-source image fusion algorithm based on transform domain and done some improvements to the existing image fusion algorithms.Two improved multi-source image fusion algorithms have been proposed in this paper which are mainly applied in the field of multi-focus image fusion.Firstly,a multi-focus image fusion algorithm based on adaptive pulse coupled neural networks using nonsubsampled contourlet transform(NSCT)is proposed in this paper.The proposed method overcomes the problem of the loss of fusion information and the poor robustness of the algorithm in the NSCT method,and gets better fusion effect.We compared the proposed method with other four methods.We can conclude that the proposed method performs better on both subjective and objective evaluation indicators.Secondly,a method of image fusion based on the combination of multi-scale decomposition and improved sparse representation is proposed.The proposed method can combine the advantages of sparse representation and multi-scale decomposition.It can solve the problem of decomposition level as well as low contrast of transform base image fusion.As a result,the proposed method can better preserve the information in the source images.A whole image is used in the process of sparse decomposition but not blocks of image,which can raise the efficiency of sparse decomposition.Results show that the proposed method performed in both the objective and subjective index.