Research On Contourlet Transform And Application In Image Processing

Contourlet transform present anisotropy of image with different multi-directions in each scale, and it can better capture localization of image or the geometry of image edges than wavelet in Image Processing. Image Fusion is integrating two or more sensors at the same time or at different times of image or video sequence to generate a new interpretation of this scene. Its main purpose is increasing reliability or image resolution by reducing uncertainty through redundancy of different images. Image Retrieval is finding out inquiries related images or similar images in library based on had extracted image features automatically.The main contents of the study and the results are as follows:1,Research on the basic content of the image fusion. In this paper the basic approach of Image Fusion was studied and focused on the wavelet-based image fusion. Also some basic theory of the image retrieval was given.2,The theory of multi-scale geometric analysis and Contourlet Transform was researched. Experiments showed contrast the multi-direction and anisotropy between contourlet and wavelet. Expansion of Contourlet Transform was also be researched.3,The Hidden Markov Model (HMM) was studied which can accurately describing the statistical characteristic of coefficient. Research also included the wavelet domain Hidden Markov Model algorithm.4,Contrast Contourlet Transform and Wavelet Transform. Image fusion using Contourlet Transform was researched which compared with the existing method for image fusion. And a tool box of image fusion was given.5,Experiment for the texture image retrieval by Contourlet-HMM model was achieved which included searching One-way or multi-direction texture image.