The Application On Image Processing On Two-dimensional Empirical Mode Decomposition And Improved Methods

In recent years, Empirical mode decomposition is proposed a new signal processing method,it is adaptive decomposition process that completely driven by the data itself, it does not depend on pre-set basis function.So It can effectively non-stationary nonlinear signal analysis. Its adaptability and good local performance of signal instantaneous frequency and amplitude, thus make the signal expressed as a real physical form. As the magority of reality signals is non-stationary and nonlinear signals, empirical mode decomposition and its application has become a hot research topic at home and abroad in recent years. This paper analyzes the principle of empirical mode decomposition method, and it analyzes key technologies and proposed an improved algorithm.In this paper, one-dimensional (EMD) and two-dimensional (BEMD) empirical mode decomposition analysis of algorithms is studied and to do a comparative study of two techniques between achieving BEMD extreme points in the extraction and decomposition of envelope surface fitting to. The field of comparative way is proposed to extract maximum points, method based on radial basis function to obtain envelope decomposition of the image. Examples show that this method can decompose the image frequency for gradually reduced N-and IMF and the remainder.Analyzed the BEMD theory and decomposition process, summarized the problems of BEMD. "Edge Effect", the interpolation function selection, screening termination condition selection, calculation amount is all the classic problem of BEMD. This paper presents a BEMD improved method, it abandons the use of interpolation algorithm to determine the envelope function, in order to avoid the complex process of "edge effect" and termination condition selection, reduce computational times or even several hundreds of times, solve the problem that BEMD needs long running time from the fundamental. It simplifies the decomposition algorithm, precides decomposition effects, speeds of Decomp- osition rate, and widens the range of applications.By the research of the nature and image characteristics of BEMD, improved BEMD method will be applied to image edge extraction and denoising. Edge information contained in the high frequency part, while the first IMF which decomposed by BEMD is high-frequency components of the reflected image, it uses its features and detect edge information of the image. Noise information is also included in the high-frequency part of the image. the second BEMD decomposition method can remove image noise, increase image visual effect. The example shows that the BEMD applied to the image edge detection and denoising can be very good.