Multiple Focus On Image Fusion Based On Pcnn And Local Variance Theory And Algorithms

Multi-focus image fusion is to fuse several focus position different source images to get a new focused image through certain fusion algorithm. In recent years, a lot of image fusion algorithms are presented such as traditional modulation method, the maximum and minimum selected method, weighted average algorithm, wavelet transformation method, pyramid method and PCA method, and so on. But from the image evaluation perspective, there are still a lot of problems.Pulse coupled neural network (PCNN) is the third generation artificial neural network with the biological background. PCNN plays a very important role in image processing. It is mainly used for image segmentation, image smoothing, edge recognition, and so on. In the image fusion, PCNN model also plays a very important role.Local variance is to point to divide an image into image block, namely into several small matrix, corresponding to each matrix will have a variance, it is called local variance. The local variance of the image can well reflect the detail of the image information and high frequency component.In this master’s thesis, the principle and implementation methods of PCNN model method are analyzed. And applications of the local variance in image processing are discussed in detail. Finally, this master’s thesis will PCNN model combining the traditional modulation method, the characteristics of the local variance get a new image fusion algorithm. This master’s thesis focused on the following:1. Detailedly analyze the principle and method of PCNN model. And introduce the advantages of PCNN model in image processing.2. Introduce the definition and significance of image fusion. Make simulations of the popular image fusion methods, and then evaluate every method according to the results of simulation.3. Through the study of local variance, this master’s thesis gets a new image fusion algorithm combining the traditional modulation method. Through experiment simulation and comparison of the popular fusion algorithm, this master’s thesis presents the image fusion algorithm whether in the subjective or objective aspect is one kind of effective and simple fusion algorithm.