Dissertation > Excellent graduate degree dissertation topics show

Multi-scale pulse coupled neural networks for image fusion

Author: WangHaiXu
Tutor: LiLei
School: University of Electronic Science and Technology
Course: Signal and Information Processing
Keywords: Image Fusion Nonsubsampled contourlet transform Wing Chong coupled neural network Scharr operator Local spatial frequency
CLC: TP391.41
Type: Master's thesis
Year: 2011
Downloads: 83
Quote: 1
Read: Download Dissertation

Abstract


Image fusion is meant to be consistent with human visual perception of integrated images. Currently, the wavelet transform has good vision due to multi-scale and time-frequency localization features, and is widely used in the field of image fusion. But for the two-dimensional image data structure occupies a very important position in the lines, curves, and other details of the edge direction information, the wavelet transform can not be effectively represented. Multi-scale geometric analysis downsampling contourlet transform Africa (NonsubsaJnpled contourlet transfbrm, NSCT) is called a real image is the most sparse representation. It inherits the properties of wavelet multi-scale and time-frequency locality, there are multi-directional and anisotropy, while addressing the contourlet transform as a result of sampling translational invariance of the ringing effect. Wing Chong Coupled Neural Network (Pulse.coupled Neural'Networks, PcNN) Because of mammalian visual system neuronal information processing and signal transduction principle be applied to the field of image fusion. Therefore, the multi-scales' NSC'T be combined with PcNN fusion method has greater value, but also one of the focuses of this study. This paper studies the multi-scale PcNN image fusion method, which reads as follows: 1, the wavelet transform, NSCT principle, the filter structure and design is described in detail, and summarizes various features of their application with the physical significance. The principle on PcNN model, parameter setting and operation mode is described in detail, combined with the objectives and characteristics of image fusion extensive experiments and discussion. 2, multi-scale decomposition of wavelet fusion method shortcoming that the wavelet coefficients improved local Scharr fusion method, since SchalT operator has more features and optimal rotation invariant approach, experiments show that this method has the same advantages and robustness. 3, due to the traditional multi-scale decomposition and PcNN based on a combination of multi-scale method does not consider the effect of the field coefficients, so that the image edge detail decreased sharpness, while taking advantage of PcNN ignition maximum rule to select appropriate factor, causing boundary effects. According to the local spatial frequency can be a proper characterization of the characteristics of image edge details, improved fusion rules effectively eliminate the boundary effect, proposed NSCT improved spatial frequency domain PcNN image fusion method (NSCTSFPCNN), and proved its effectiveness. 4, multi-scale image fusion PcNN model under the framework of various methods step experiments and discussion, comparisons of this framework NSCTSFPCNN method has the advantage, then the key parameters of the method for each of five experiments and discussions, the optimal principles and parameters set range.

Related Dissertations

  1. Research on Joint Target Detection for Dual-Sensor Image and System Implementation,TP391.41
  2. The Research of Application of Ridgelet Transform in the Fusion of Multispectral and Panchromatic Images,TP391.41
  3. Image Stitching technology research,TP391.41
  4. Remote Sensing Image Based on Genetic Algorithm Fusion,TP751
  5. Research on Remote Sensing Image Fusion Algorithm Based on Wavelet Transform,TP751
  6. Research of Applications of Remote Sensing Image Fusion in the Land-use Change Monitoring,P237
  7. Quality Detection of Fruit by Infrared Image Fusion Based on DSP,TP391.41
  8. The Fusion Method of Digital Images Based on Multiresolution Analysis Theory,TP391.41
  9. Recearch on Applications of Contourlet Transform in Image Processing,TP391.41
  10. Research on Digital Image Fusion Algorithms,TP391.41
  11. Study of Image Fusion Based on Curvelet Transform,TP391.41
  12. The Study of Remote Sensing Image Data Fusion Methods and the Effect of Evaluation,P237
  13. The Research of Feature Selection and Segmentation of Medical Image Based on Fusion,TP391.41
  14. Research on Digital Image Watermarking Algorithm Based on Naturalness Preserving Transform,TP309.7
  15. The Research on Multi-focus Image Fusion,TP391.41
  16. Image Retrieval Based on Color and Non-Subsampled Contourilet Features,TP391.41
  17. Research on Multisensor Image Fusion Algorithoms,TP391.41
  18. Study of Image Fusion Based on Contourlet Transform,TP391.41
  19. Methods on Medical Image Fusion Based on DICOM Standard,TP391.41
  20. Research and Application of Finish Image Based on NSCT,TP391.41
  21. Fusion Technology of Infrared and Visible Images,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