Study On Multi-resolution Characteristics Of Image

Multi-resolution methods, as it can provide different resolution information of image, integrates the accuracy of high-resolution with the easy severability of low-resolution, and controls resolution very easily. So the multi-resolution remote sensing image processing approach has been more in favor of the researchers. Today image resolution and image data have been increasing rapidly. For specific occasions, how to find the appropriate resolution in which we can do segmentation of the interest region, target recognition, reducing the data of image, and improve the performance of real-time processing system better. Multi-resolution methods become the inevitable means to solve those problems.This paper firstly introduces the background and development of multi-resolution image analysis methods. It mainly focuses on the necessity of processing image in different resolution, and proposes the key to finding the best resolution of the image which depending on the relationship between the features and resolution of image. Then it analyzes the existing multi-resolution analysis techniques and advances a method of pixel mapping between different resolutions based on image pyramid to solve the devastation of important characteristics of image caused by decomposition of image pyramid. After the multi-resolution model is set up, the relationship between resolution and image features are explored to establish the uncertainty theory of resolution on image features. Those image features include statistical characteristics, fractal characteristics, geometric characteristics, information characteristics, moments characteristics, etc. Finally, an efficient image segmentation method for complexity image is presented, which is developed based on the evolution of local variance in multi-resolution images for different objects. It enlarges the difference in background and goals for image segmentation. In this paper, the idea of multi-resolution analysis for images has certain innovations and achieves some results.In addition, the experimental results show excellent performance of the image segmentation method proposed in this paper.