Docstoc

Image enhancement of UV Solar Images

Document Sample
Image enhancement of UV Solar Images Powered By Docstoc
					                                                                              Ron Caplan
                                                                                Math 336
                                                                       Final Presentation
                                                                      December 8th, 2008




Images courtesy of SOHO/[instrument] consortium. SOHO is a project of international cooperation between ESA and NASA. where
[instrument] stands for the name of the instrument that acquired the data
Overview
   Introduction
   RAW Data
   Pre-processing
     Format
     Degridding
 Stationary Wavelet
  Transform
 Results
     My results
     Paper’s Results
   Conclusion
    Introduction


 SOHO, the Solar and Heliospheric Observatory, is a project of international
  cooperation between ESA and NASA to study the Sun, from its deep core to the
  outer corona, and the solar wind.
 One of its many instruments is the Extreme ultraviolet Imaging Telescope (EIT),
  which provides full disc images of the Sun at four selected colors in the extreme
  ultraviolet (EUV), mapping the plasma in the low corona and transition region at
  temperatures between 80,000°C and 500,000°C.
 The EIT can image active regions, filaments and prominences, coronal holes,
  coronal "bright points," polar plumes, loops, and arcades, as well as dynamical
  events such as flares and mass ejections.
 However, the multiscale nature of the observed solar features has not been
  fully exploited so far. Guillermo Stenborg, Angelos Vourlidas, and Russell A.
  Howard have come up with a wavelet-based processing technique that
  enhances the EUV images based on their multiscale nature, and reveals
  features not seen with standard image-processing techniques . They have
  processed the entire EIT data set with their technique, and has made it
  available to solar physicists.
    Raw
    Data
 uint16 Format
  (0 -> 65,535)
 imagesc(im)
 imadjust(im)



    Issues
 Grid
  patterns
 Noise
          Pre-Processing
   Rescale RAW data extremes:
      Pre-Processing
   Rescale with imadjust() and
    convert to uint8 grayscale:
Pre-Processing
   De-grid using manual threshold-based notch filter on
    fft2 of image:
   Original
   My De-Grid
   Official
    Image
Stationary Wavelet Transform
 Kernal (filter) B3-Spline / Biorthogonal 3.3
 Upsample filter at each level (pad with zeros)
 Coefficient Matrix 2Nx2N at each level
 Set weights for each levels detail coeffs to bring out structure
 Also brings out more noise/grid – paper does more processing
 MATLAB: swt2.m/iswt2.m
    My
    results
   Using 'bior3.3‘,
    J=5
    w=[2 5 4 3 2]:


    Using official
    image:
    My
    results
   Using 'bior3.3‘,
    J=5
    w=[2 5 4 3 2]:



    Using my
    degrid:
    Results
    from
    Paper
 2-Stage
  Process     Residual Light Model   Noise Mask
 Original
  image
 Processed
  image
Conclusion

 Process works very well to bring out
  features not apparent in the images.
 Technique is not as simple as it
  appears…

				
DOCUMENT INFO
Shared By:
Categories:
Tags:
Stats:
views:3
posted:7/15/2011
language:English
pages:13