원문정보
초록
영어
In order to solve the problem that the present logging image reconstruction algorithm has certain staircase effect and ring effect during the discrete Poisson noise image repair process, a noise reduction algorithm based on Schatten matrix operator for discrete Poisson logging image is proposed in this article. Specifically, Schatten norm matrix of Hessian operator is adopted to estimate each pixel value of the image, and meanwhile the non-quadratic regularization function is adopted to eliminate the staircase effect; then Poisson measurement model is integrated with Schatten regularization function to construct the minimum objective function; then norm proximal mapping estimation algorithm is defined and is adopted to decouple the data fidelity item and the regularization matrix in the objective function, thus to recover the image. The simulation result shows: compared with present denoising technology, the algorithm proposed in this article has better repair effect and greater similarity in the aspect of processing the discrete Poisson noise.
목차
1. Introduction
2. Degradation Model
3. Matrix Operator Regularization
3.1. Sp Norm Matrix
3.2. Regularization of Hessian Operator and Matrix Operator
4. Poisson Image Reconstruction Algorithm Design
4.1. Punishment ML Estimation Value
4.2. Decoupling of Data Fidelity Item and Regularization Matrix
5. Enclosure Form of Approximate Mapping of Sp Norm ( p = 1,2,∞)
6. Simulation Result and Analysis
6.1. Noise Reduction Effect of Discrete Poisson Logging Image
6.2. Contrastive Analysis of Structure Similarity SSIM
7. Conclusion
Acknowledgement
References