Image Processing Of Compressed Sensing Based On EMD Method

Empirical mode deposition (EMD)proposed is a new multi-scale analysis method, which has good performance especially in the non-linear non-stationary signal analysis. Compared to the classical Fourier transform, advantage of EMD is non-stationary signal analysis. EMD used as a one-dimensional signal processing is relatively mature and widely, and has good performance, so foreign scholars expanded it to two-dimensional, proposed a two-dimensional empirical mode decomposition method and applied to two-dimensional signal processing, this also acquired well performance.In recent years the rise of compressed sensing (Compressive Sensing, referred to as CS) signal processing theory is another area of outstanding achievements. It features only a small amount of data collected will be able to reconstruct the original data, breaking the laws of the Nyquist sampling limit. This mitigate the pressure in real-world applications which need to a large amount of data.This study carried out two aspects:First, the theoretical simulation. Made a detailed description of EMD and CS theory, from one-dimensional extended to two-dimensional. Proposed the CS-dimensional image compression algorithm based on EMD method, then simulated it and done comparison tests. Second, designed a simple web simulation application based on the algorithm simulation. The main work of this paper is as follows:Made a detailed introduction to the theory of CS and EMD algorithm, given a specific algorithm implementation steps, simulated and verified the feasibility from the one-dimensional to two-dimensional.According to the basic theories of CS and BEMD, this paper attempts to proceed image processing combining the BEMD and CS: first, sparsed the image using Anna’s method, then compressed using CS, at last using compressed data to restore the original imageDesigned web-based simulation system for remote image compression using Java builder tool provided by Matlab, making full use of Matlab’s powerful data processing capabilities. The simulation system used image processing method fusing CS and EMD proposed in this paper.