The Applications Of Compressive Sensing In Image Denoising

The technology of digital image processing is widely applied in people’s daily production and life, which is the hotspot in research of technology. Image processing is rich in content, including image denoising which is one of the most critical technology. Now the related research of image denoising is more and more widely and it will be the development of this field with using the new technology.Compressive sensing is emerging as a data sampling technology in recent years, the core idea of which is taking the fewer sample data to reconstruct signal. Compressive sensing breaks the traditional sampling methods, which uses the sparse representation of signals to ensure the structure of the original signals, then to refactor the original signals accurately through the algorithm.The contents of this paper are the applications of compressive sensing in image denoising. According to the compressive sensing technology we establish the image model with noise for removing the noise through computing the sparse representation of image and reconstruct the original image by algorithm. The main work of this paper as follows:The knowledge of compressive sensing is expounded, including its main structure and reconstruction algorithm.The knowledge of image denoising is expounded, including digital image^image quality evaluation image noise and the traditional denoising methods.The stability of sparse solution is to analysis in compressive sensing, and the applicable stability of RIP conditions is to analysis. It is to achieve the applications of compressive sensing in image denoising. The sparse representation of image is obtained across the discrete wavelet transform rather than the traditional discrete Fourer transform. The experiments proved that the method has the feasibility and effectiveness. It is to achieve the new reconstruction algorithm by improving the OMP method and combining the coordinate descent method, which reduces the complexity of the calculation. The experiments show that this method is feasible and effective better than the prior denoising method.When solving the sparse representation of the image, it is to apply the redundant dictionary instead of the traditional linear orthogonal transformation. According to compressive sensing, it is to receive the image denoising using the K-SVD algorithm. The experimental results show that the effect is good. According to the image translation invariance principle and changing in the update state of the dictionary learning, it is to improve the K-SVD algorithm. Then it is to apply the new algorithm into the image denoising based on the compressive sensing. The experimental results show that this method has the feasibility and effectiveness.