Dissertation > Excellent graduate degree dissertation topics show

Research on Image Denoising Approach Based on Wavelet and Its Statistical Characteristics

Author: HouJianHua
Tutor: TianJinWen
School: Huazhong University of Science and Technology
Course: Control Science and Engineering
Keywords: Wavelet Transform Denoising Wavelet threshold denoising Wavelet domain Wiener filtering Statistical model Statistical properties of the wavelet coefficients Neighborhood correlation Bayesian methods
CLC: TP391.41
Type: PhD thesis
Year: 2007
Downloads: 2308
Quote: 22
Read: Download Dissertation

Abstract


The image in the process of access or transmit inevitably be affected by noise pollution, noise in the image seriously affected the subsequent image processing, such as image segmentation, coding, feature extraction and target detection. In order to improve the quality of the image as well as follow-up to a higher level of processing needs for image denoising on to become the image preprocessing a very important work. Image denoising purpose is to recover from the noisy image noise pollution \The traditional image denoising method unsatisfactory compromise in noise reduction and insurance details; wavelet transform as a new time-frequency analysis method, multi-scale, multi-resolution analysis of the characteristics of the signal processing provides a new powerful means, has been successfully applied in the field of image denoising. Wavelet-based denoising method has become a major branch of image denoising and restoration, according to the statistical properties of the image wavelet coefficients, to study the model-based de-noising method, the main research directions in the field of image denoising, both in theoretically or in practical applications has important significance. Theory of wavelet analysis tools, image the denoising theory and method of wavelet domain system, in-depth research, the main work includes the following four parts: 1, the first two chapters of the wavelet image denoising method research summarized in this article as the basis of the full text. a comprehensive summary of the research of wavelet-based image denoising method. The First image denoising technology development, in particular, the method of wavelet image denoising progress. Yet a more comprehensive classification method for wavelet image denoising field, the three stages of development of the field as a clue, wavelet image denoising algorithm is a new classification and is divided into four categories, and each type in the representation of the algorithm discussed. Threshold denoising is a very important method in the study of wavelet denoising, this system, in-depth analysis on the core issue of threshold selection and combination of specific algorithm in the principles and methods on the threshold of the most representative on clear, and these typical algorithms respectively orthogonal wavelet transform base and translational invariant wavelet basis under the comprehensive experimental simulation and analysis discussion. Get some meaningful conclusions from the experimental results and performance analysis. Plus-noise signal generation in one-dimensional signal denoising algorithm simulation method, to produce high-precision signal-to-noise ratio of plus-noise signal on the basis of rigorous theoretical derivation. 2 Wavelet domain Wiener filtering method in wavelet domain Wiener filtering method is a very active area of ??research in the field of wavelet image denoising content. This paper defines the three forms of the wavelet domain Wiener filtering three new denoising algorithm. Iterative Wiener filtering algorithm of wavelet domain is proposed; in wavelet domain experience Wiener filter based on with BayesShrink threshold algorithm to improve the accuracy of the estimate of the desired signal, while taking advantage of the multi-wavelet basis to better capture certain characteristics of the signal , and to achieve the iteration, and thus significantly enhance the denoising performance of the algorithm. A new image combination filter; first BayesShrink algorithm of image preprocessing, then the airspace Lee filter; the algorithm core of an estimated pre-denoising variance of the residual noise in the image near optimal formula, in order to ensure a match between the two algorithms. A wavelet domain local adaptive image denoising algorithm; theoretical analysis, the estimation error of LAWML algorithm the local variance estimation of an observation coefficient threshold compared, and LAWML algorithm, the new algorithm in objective peak signal-to-noise ratio and subjective visual effects there are significant improvements. 3 study image denoising method based on a statistical model of the wavelet coefficients wavelet domain denoising based on statistical models Bayes-depth study of improvements for existing deficiencies of the two algorithms: Sendur bivariate model denoising algorithm using MAP soft threshold of the three most high-frequency sub-band local adaptive processing; Moulin's with different edge standards based the Laplace model MapShrink subband adaptive algorithms, wavelet coefficients modeling Laplace distribution model parameters are estimated using the local neighborhood window, the making MapShrink threshold having a local self-adaptive. The introduction of wavelet coefficients classification techniques to image denoising, proposes two new algorithms. First Gaussian mixture model based on a Gaussian mixture model with a pixel adaptive adjustment, using local Bayesian classification threshold of wavelet coefficients, through the the current coefficient neighborhood windows in the two categories coefficient information, the model parameters estimate; to design Wiener filter based on the MMSE criterion. The second method is the one-dimensional signal wavelet neighborhood threshold extended and applied to a two-dimensional image, each wavelet coefficient in the sub-band based on the threshold value of the size of its neighborhood is divided into \; \The experimental results show that the algorithm has a low computational complexity and denoising performance features. Wavelet statistical model-based SAR image speckle suppression method as specific examples of image denoising, the end of this article discusses the application of wavelet denoising SAR image speckle noise suppression. Recalling the SAR image speckle suppression method progress, focus on the SAR image speckle method based on wavelet statistical model is proposed based on Bayesian MAP estimate wavelet domain local adaptive speckle algorithm. Containing plaques image do Bayesian estimated number transform and redundant wavelet decomposition, speckle noise, the useful signal wavelet coefficients are modeled as a Rayleigh distribution, Laplace distribution, the MAP criterion to get a parse expression style, and prove its processing in essence, is a kind of soft threshold denoising, and thus has the characteristics of the algorithm is simple; further Laplacian parameter is estimated by the local neighborhood window algorithm with local adaptivity. The theoretical analysis and experimental simulation show that the algorithm can effectively suppress speckle noise of SAR images better maintain the image of the strong, weak detail.

Related Dissertations

  1. Research on Electromagnetic Acoustic Transducer Detection System Based on FPGA,TH878.2
  2. Research on Methods of Medical Ultrasound Image Denoising,TP391.41
  3. Feature Extraction, Selection and Combination in Lipreading,TP391.41
  4. Design of Testing Equipment for Electronic Products,TN06
  5. Research on Identification System of Cashmere and Wool Fiber,TS101.921
  6. Characteristics of sensory stimulation evoked,R318.0
  7. Network transmission ROI image coding algorithm,TN919.81
  8. Image Fusion Algorithms Based on Multi-scale Analysis,TP391.41
  9. Feature Extraction Technologies Research and Implementation of 3D Models Based on Wavelet Transform,TP391.41
  10. Research on Contourlet Transform and Its Application on Image Processing,TP391.41
  11. An Algorithm on Clustering and Anomaly Detection for Multiple Data Streams,TP311.13
  12. Research on Information Hiding Based on Images,TP309.7
  13. Design and Implication of Image Compression System Based on Wavelet and DM6446,TP391.41
  14. Combined DWT Dynamic Data Reconciliation Research and Application,TP274
  15. Traffic Prediction Model Based on Wavelet and Markov,TP393.06
  16. Wavelets-based Signal Sparse Representation and Its Application in Image Denoising,TP391.41
  17. The Study of Image Watermark Base on Discrete Wavelet Transformation,TP309.7
  18. The Compression and Application of the Brain CT Image,TP391.41
  19. Research on Methods of Image Processing of the Image Information Processor,TP391.41
  20. Inspection and Recognition on a Class of Tool Processing Trace,TP391.41
  21. Research on Image Denoising Algorithm in Machine Vision System,TP391.41

CLC: > Industrial Technology > Automation technology,computer technology > Computing technology,computer technology > Computer applications > Information processing (information processing) > Pattern Recognition and devices > Image recognition device
© 2012 www.DissertationTopic.Net  Mobile