Dissertation > Excellent graduate degree dissertation topics show

Image Denoising Research Based on Wavelet Transform

Author: PanYang
Tutor: LiWenHui；ZhangZhenHua
School: Jilin University
Course: Software Engineering
Keywords: Denoising Wavelet Transform Threshold
CLC: TP391.41
Type: Master's thesis
Year: 2010
Downloads: 408
Quote: 0
Read: Download Dissertation

Abstract


Generally, the image acquired by the various ways in reality, a certain degree of noise pollution, thereby bring our everyday applications and analysis difficult. The image denoising object to reduce and eliminate the noise in the image, thereby improving image quality. In recent years, wavelet has received much attention in the areas of mathematics, signal processing and image processing and many scholars have launched a large number of detailed studies of wavelet-based image denoising method. First proposed by Donoho wavelet thresholding is simple, effective become a mainstream method. There are a variety of ways, so far, there are still many scholars in the study on the basis of its. However, due to the natural design defects wavelet thresholding method for the removal of noise and retain edge details, that is not possible to present a threshold method can perfect noise and image information separately, so the traditional threshold denoising method can not be at the same time to achieve de-noising and maintain the edge information serves two purposes. The purpose of image edge detection method is effective edge information of the detected image, while inhibition we do not need, such as noise and other unwanted information. Traditional edge detection operators can more effectively detect the edges of the image, but very sensitive to noise image. In wavelet domain wavelet coefficient correlation edge detection is an effective noise suppression method. Take into account the different circumstances of the noise point with normal attenuation coefficient scale transformation, the correlation coefficients can effectively detect the edge information while suppressing noise. However, the method is often not as good as the traditional edge detection operator detection effect, taking into account the scale transformation due to the coefficient offset. The paper is organized as follows: In the first chapter, we first introduce the topics of our background significance, as well as the development of image denoising Denoising evaluation criteria, and finally a brief course of development of wavelet image denoising method. The second chapter we work, where we mainly introduce the basic theory of the wavelet transform two-dimensional wavelet transform and the lifting wavelet theory, at the end through the literature, we analyzed the noise point and normal wavelet transform coefficient distribution law. The third chapter introduces the basic the wavelet denoising theory and several. First, we introduce a threshold selection method, including the hard threshold, soft threshold, semi-soft threshold. Secondly, we introduce several threshold criteria. Such as the law of universal threshold, minimax threshold, Stein unbiased risk threshold, Bayesian threshold. In the fourth chapter, we first introduced the Bayesian threshold algorithm, including how to estimate the variance of the noise point has been how to calculate the threshold - point-by-point threshold method and block-by-block threshold method. Chapter V, the first wavelet-based edge detection methods. Which mainly include traditional edge operator method and wavelet coefficients of correlation algorithm. Both algorithms, each with its own characteristics. Raised our improved method. Effectively two algorithms, thus gives our algorithm to detect the edge information. Finally, we given algorithm based on a large number of experiments, experiments show that our algorithm can extract image edge, the edge effectively suppress noise, denoising effect, our algorithm is not only The denoising effective, better reflect the edge. PSNR, our algorithm than the traditional Bayes threshold denoising method and class zerotree Bayesian threshold denoising method, there is some improvement. Algorithm denoising results from the theoretical and experimental results have proved that, relative to traditional Bayes threshold method and improved Bayesian threshold denoising method, we give a stable, fast, and at the same time more effective.

Related Dissertations

  1. Research on Electromagnetic Acoustic Transducer Detection System Based on FPGA,TH878.2
  2. Design of Testing Equipment for Electronic Products,TN06
  3. Research on Identification System of Cashmere and Wool Fiber,TS101.921
  4. Battle tanks in tank bottom detection technology,TE972
  5. Methods based on RDT FDF diagnostic systems for oil drilling accident,TE28
  6. H.264 fast intra prediction mode selection algorithm,TN919.81
  7. Hypercomplex sparse representation and its applications,TP391.41
  8. Geometrically Robust Image Watermarking Algorithm,TP309.7
  9. Image fusion method based on variational partial differential equations,TP391.41
  10. Web optimization methods for monitoring of sensitive information,TP393.08
  11. Research on Cutting Tool Wear Detection Based on Computer Vision,TP274
  12. Marx from the Perspective of Globalization and the Construction of Universal Values,A811
  13. Comparison of regional differences of the household registration reform,D631.42
  14. The Otherness from the Perspective of Chinese and Western Culture,I712
  15. Research of the Relevant Sensitized Acupoints on the Body Surface in Chronic Superifical Gastritis,R246
  16. Software Technology Research of Real-time Arrhythmia Monitoring System Based on Smart-phone,R541.7
  17. Research on the Influence of 3d Electric Field on the Excitability of the Optic Nerve,R318.0
  18. Study on Engineering Drawings Vectorization,TP391.41
  19. Multi-spectral remote sensing image registration and fusion research,TP751
  20. Signal Processing Method Based on Singular Value Decomposition and Its Application to Mechanical Fault Diagnosis,TH165.3
  21. The Research and Design of Ultrasonic Non Destrucive Testing Equipment,TH878.2

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