SIRW October 23-24, 2003 Pattern Recognition Pattern Recognition Techniques Applied to Solar Image Data Techniques Applied to Solar Image Data A wavelet packets A wavelet packets equalization technique equalization technique to reveal the multiple to reveal the multiple spatial-scale nature spatial-scale nature of coronal structures of coronal structures Guillermo A. Stenborg Guillermo A. Stenborg The Catholic University of America & NASA Goddard Space Flight Center
63
Embed
Pattern Recognition Techniques Applied to Solar Image Data SIRW October 23-24, 2003 Pattern Recognition Techniques Applied to Solar Image Data A wavelet.
This document is posted to help you gain knowledge. Please leave a comment to let me know what you think about it! Share it to your friends and learn new things together.
Transcript
SIRW October 23-24, 2003 Pattern Recognition Techniques Applied to Solar Image DataPattern Recognition Techniques Applied to Solar Image Data
A wavelet packets A wavelet packets equalization technique equalization technique to reveal the multiple to reveal the multiple spatial-scale nature spatial-scale nature of coronal structuresof coronal structures
Guillermo A. Stenborg Guillermo A. Stenborg
The Catholic University of America&
NASA Goddard Space Flight Center
SIRW October 23-24, 2003 Pattern Recognition Techniques Applied to Solar Image DataPattern Recognition Techniques Applied to Solar Image Data
ObjectiveObjective
• More accurate tracking of coronal events
• More accurate determination of onset times
• Tracking of continuous coronal outflow (slow solar wind ?) seen in LASCO-C2 and -C3 images ?
ApproachApproach
Selective contrast enhancement of boundaries and internal details of
coronal features
More reliable identification of coronal structures to More reliable identification of coronal structures to help in the process of automatic recognition of help in the process of automatic recognition of coronal eventscoronal events
SIRW October 23-24, 2003 Pattern Recognition Techniques Applied to Solar Image DataPattern Recognition Techniques Applied to Solar Image Data
WTs (Wavelet Transforms)
• Transforms data to time-scale domain
• Use of “mother wavelets”
How do we analyse signals?
• Dilations and compressions
• Traslations over the signal´s domain
Analyzing wavelet adapted to frequency
Spatialy Localized
Additional capabilities & features
• Infinite set of possible basis functions
• Quantitative measure of information
Adapted wavelets
Time-scale based methodsTime-scale based methods
SIRW October 23-24, 2003 Pattern Recognition Techniques Applied to Solar Image DataPattern Recognition Techniques Applied to Solar Image Data
1) The technique consists in decomposing a given signal in the so-called wavelet scales or wavelet planes, the the
first scalesfirst scales containing the higher (spatial) frequency componentsthe higher (spatial) frequency components and the last onesthe last ones containing the lower (spatial) the lower (spatial)
frequency signaturesfrequency signatures.
2) Wavelet Transform properties allow further decomposition of each wavelet scale in subsequent scales.
3) After noise filtering in the wavelet domain, and assigning different weights to the wavelet scales (including a smoothed array
called “continuum”) a reconstructed image is obtained, showing selectively contrast- enhanced features (in a way
resembling the technique known as „unsharp masking“).
The Wavelet-based Equalization TechniqueThe Wavelet-based Equalization Technique
SIRW October 23-24, 2003 Pattern Recognition Techniques Applied to Solar Image DataPattern Recognition Techniques Applied to Solar Image Data
The 1D “à trous” algorithm
Bn-splines (1D)Mother
Wavelets
Analysis produces a set of resolution-related views of the original signal, called scales.
Scaling is achieved by dilating and contracting the basic wavelet to form a set of wavelet functions.
Wavelet ScalesStarck J.-L. et al., ApJ, 1997
Wavelet Transform
MW: B3-spline (1D)
SIRW October 23-24, 2003 Pattern Recognition Techniques Applied to Solar Image DataPattern Recognition Techniques Applied to Solar Image Data
The 2D “à trous” algorithm
Weight
0 1
1 1
2 0
3 0
4 0
5 0
SIRW October 23-24, 2003 Pattern Recognition Techniques Applied to Solar Image DataPattern Recognition Techniques Applied to Solar Image Data
The 2D “à trous” algorithm
Weight
0 1
1 0
2 1
3 0
4 0
5 0
SIRW October 23-24, 2003 Pattern Recognition Techniques Applied to Solar Image DataPattern Recognition Techniques Applied to Solar Image Data
The 2D “à trous” algorithm
Weight
0 1
1 0
2 0
3 1
4 0
5 0
SIRW October 23-24, 2003 Pattern Recognition Techniques Applied to Solar Image DataPattern Recognition Techniques Applied to Solar Image Data
The 2D “à trous” algorithm
Weight
0 1
1 0
2 0
3 0
4 1
5 0
SIRW October 23-24, 2003 Pattern Recognition Techniques Applied to Solar Image DataPattern Recognition Techniques Applied to Solar Image Data
Weight
0 1
1 0
2 0
3 0
4 0
5 1
The 2D “à trous” algorithm
SIRW October 23-24, 2003 Pattern Recognition Techniques Applied to Solar Image DataPattern Recognition Techniques Applied to Solar Image Data
The 2D “à trous” algorithm
p
jj yxwyxwyxI
10 ),(),(),(1) Reconstruction
of original image
),(1 yxw ),(2 yxw ),(3 yxw ),(4 yxw
),(0 yxw
),( yxI
p
jjj yxWyxWyxI
10 ),(),(),(' 4) Weighted Reconstruction:
The Figure depicts the wavelet scales 1 to 4 (Wi) and the smoothed image (Wo), i.e., continuum, of I (x,y)
= EIT Fe IX/X ( 171 Å) image = =1 =[1,1,1,1]k=0
),(' yxI
For comparison, continuum corresponding to decomposition
based on 50 scales
when
2DB3-spline
:
:
),(ˆ),( yxyx Ijj 2) Local standard deviation of Noise at scale j
Local standard deviation of noise in original image (first scale)Noise progression in wavelet space
),(),( if ),(
),(),( if 0),(
yxkyxwyxw
yxkyxwyxW
jjj
jj
j
3) Noise filtering: (hard thresholding as in Donoho & Johnstone, 1994)
SIRW October 23-24, 2003 Pattern Recognition Techniques Applied to Solar Image DataPattern Recognition Techniques Applied to Solar Image Data
The 2D “à trous” algorithm
ReconstructionWeight
0 1
1 5
2 5
3 5
4 5
5 5
SIRW October 23-24, 2003 Pattern Recognition Techniques Applied to Solar Image DataPattern Recognition Techniques Applied to Solar Image Data
• The technique shown so far involves decomposing a given image in wavelet planes (i.e., spatial frequency bands), the finer scales containing the higher frequency components and the coarser ones the lowest frequency signatures.
• For non-orthogonal wavelets (as for the “à trous” algorithm) the Signal to Noise Ratio (SNR) increases toward coarser scales.
• Straightforward filtering of wavelet coeficients at this stage produces rejection of signal along with noise.
Comments on the 2D “à trous” algorithm
A better alternative is a technique allowing a finer analysis of the frequency content of the signal
SIRW October 23-24, 2003 Pattern Recognition Techniques Applied to Solar Image DataPattern Recognition Techniques Applied to Solar Image Data
The alternative:
The Wavelet The Wavelet PacketsPackets -based Equalization Technique -based Equalization Technique • The splitting algorithm of wavelet packets on non-orthogonal wavelets allows much
better frequency localization. That is achieved by recursively decomposing (transforming) the wavelet scales obtained with the “à trous” algorithm (thanks to the fact that wavelet transform is not its own inverse).
• 1-D variant of the algorithm was first implemented for an astronomical aplication by Fligge & Solanki, 1997 to reduce noise in astronomical spectra.
• 2-D variant of the algorithm was first implemented for an astronomical application by Stenborg & Cobelli, 2003 to reveal the multiple spatial-scale nature of coronal structures (hereafter SC2003).
SIRW October 23-24, 2003 Pattern Recognition Techniques Applied to Solar Image DataPattern Recognition Techniques Applied to Solar Image Data
w0(0)
w1(0)
w2(0)
wp1(0)
...
...wk
(0) w0(0,k)
w1(0,k)
w2(0,k)
wp2(0,k)
...
...wm
(0,k)
w0(0,k,m)
w1(0,k,m)
w2(0,k,m)
wp3(0,k,m)
...
...wk
(0,k,m)
I(x,y)
The technique
Multiple-level decomposition scheme:
3-level decomposition tree. For clarity only one branch is shown at each decomposition level, but it is assumed that when computing a new level all coefficients of the previous one are decomposed
1
0
2
0
3
0
),,0(),(p
i
p
j
p
h
ijhwyxI
),( if ),(
),( if 0),(
,(...)(...)
,(...)
(...)
yxkwyxw
yxkwyxW
lkhhh
lkhh
h
),(ˆ),( (...)(...) yxyx Ihh
1)
2)
1
0
2
0
3
0
),,0(,,),('
p
i
p
j
p
h
ijhhji WyxI
3)
33
0302 01000
23
13121210 1
535251505
434241414
3231303
2221202
3210
51 p
32 pIn matrix form the weighting coefficients can be depicted as (for 2 levels):
4)
Briefly, the first level decomposition of the given image in p1 scales gives rise to the wavelet transform set {wi
(0)}, i=1...p1, i=0 corresponding to the continuum component. Afterward, further decomposition is applied to each wavelet plane.
SIRW October 23-24, 2003 Pattern Recognition Techniques Applied to Solar Image DataPattern Recognition Techniques Applied to Solar Image Data
0 1 2 3
0 1 1 1 1
1 0 1 0 0
2 0 1 0 0
3 0 1 0 0
4 0 1 0 0
5 0 1 0 0
SIRW October 23-24, 2003 Pattern Recognition Techniques Applied to Solar Image DataPattern Recognition Techniques Applied to Solar Image Data
0 1 2 3
0 1 1 1 1
1 0 0 1 0
2 0 0 1 0
3 0 0 1 0
4 0 0 1 0
5 0 0 1 0
SIRW October 23-24, 2003 Pattern Recognition Techniques Applied to Solar Image DataPattern Recognition Techniques Applied to Solar Image Data
0 1 2 3
0 1 1 1 1
1 1 0 0 0
2 1 0 0 0
3 1 0 0 0
4 1 0 0 0
5 1 0 0 0
SIRW October 23-24, 2003 Pattern Recognition Techniques Applied to Solar Image DataPattern Recognition Techniques Applied to Solar Image Data
0 1 2 3
0 1 1 1 1
1 5 15 10 7
2 3 10 7 5
3 1 7 5 3
4 1 5 3 1
5 0 3 1 1
0 1 2 3
0 1 1 1 1
1 0 0 1 3
2 1 1 3 5
3 1 3 5 7
4 3 5 7 10
5 5 7 10 15
SIRW October 23-24, 2003 Pattern Recognition Techniques Applied to Solar Image DataPattern Recognition Techniques Applied to Solar Image Data
0 1 2 3
0 0 0 0 0
1 1 0 0 1
2 1 0 0 1
3 1 0 0 1
4 1 0 0 1
5 1 0 0 1
0 1 2 3
0 0 0 0 0
1 0 1 1 0
2 0 1 1 0
3 0 1 1 0
4 0 1 1 0
5 0 1 1 0
SIRW October 23-24, 2003 Pattern Recognition Techniques Applied to Solar Image DataPattern Recognition Techniques Applied to Solar Image Data
SIRW October 23-24, 2003 Pattern Recognition Techniques Applied to Solar Image DataPattern Recognition Techniques Applied to Solar Image Data
SIRW October 23-24, 2003 Pattern Recognition Techniques Applied to Solar Image DataPattern Recognition Techniques Applied to Solar Image Data
SIRW October 23-24, 2003 Pattern Recognition Techniques Applied to Solar Image DataPattern Recognition Techniques Applied to Solar Image Data
1/5LASCO-C2: April 21, 2002
SIRW October 23-24, 2003 Pattern Recognition Techniques Applied to Solar Image DataPattern Recognition Techniques Applied to Solar Image Data
2/5LASCO-C2: April 21, 2002
SIRW October 23-24, 2003 Pattern Recognition Techniques Applied to Solar Image DataPattern Recognition Techniques Applied to Solar Image Data
3/5LASCO-C2: April 21, 2002
SIRW October 23-24, 2003 Pattern Recognition Techniques Applied to Solar Image DataPattern Recognition Techniques Applied to Solar Image Data
4/5LASCO-C2: April 21, 2002
SIRW October 23-24, 2003 Pattern Recognition Techniques Applied to Solar Image DataPattern Recognition Techniques Applied to Solar Image Data
5/5LASCO-C2: April 21, 2002
SIRW October 23-24, 2003 Pattern Recognition Techniques Applied to Solar Image DataPattern Recognition Techniques Applied to Solar Image Data
1/2LASCO-C3: April 21, 2002
SIRW October 23-24, 2003 Pattern Recognition Techniques Applied to Solar Image DataPattern Recognition Techniques Applied to Solar Image Data
2/2LASCO-C3: April 21, 2002
SIRW October 23-24, 2003 Pattern Recognition Techniques Applied to Solar Image DataPattern Recognition Techniques Applied to Solar Image Data
1998.06.02
0 1 2 3
0 0 0 0 0
1 1 1 1 1
2 1 1 1 1
... 1 1 1 1
7 1 1 1 1
8 1 1 1 1
0 1 2 3
0 0 0 0 0
1 0 1 1 1
2 0 1 1 1
... 0 1 1 1
7 0 1 1 1
8 0 1 1 1
0 1 2 3
0 1 1 1 1
1 8 8 8 8
2 8 8 8 8
... 8 8 8 8
7 8 8 8 8
8 8 8 8 8
0 1 2 3
0 1 1 1 1
1 18 8 8 8
2 18 8 8 8
... 18 8 8 8
7 18 8 8 8
8 18 8 8 8
0 1 2 3
0 1 1 1 1
1 8 18 18 18
2 8 18 18 18
... 8 18 18 18
7 8 18 18 18
8 8 18 18 18
01/14
LASCO-C2 Level 0.5
SIRW October 23-24, 2003 Pattern Recognition Techniques Applied to Solar Image DataPattern Recognition Techniques Applied to Solar Image Data
1998.06.02
0 1 2 3
0 0 0 0 0
1 1 1 1 1
2 1 1 1 1
... 1 1 1 1
7 1 1 1 1
8 1 1 1 1
0 1 2 3
0 1 1 1 1
1 8 8 8 8
2 8 8 8 8
... 8 8 8 8
7 8 8 8 8
8 8 8 8 8
0 1 2 3
0 1 1 1 1
1 8 18 18 18
2 8 18 18 18
... 8 18 18 18
7 8 18 18 18
8 8 18 18 18
0 1 2 3
0 1 1 1 1
1 18 8 8 8
2 18 8 8 8
... 18 8 8 8
7 18 8 8 8
8 18 8 8 8
0 1 2 3
0 0 0 0 0
1 0 1 1 1
2 0 1 1 1
... 0 1 1 1
7 0 1 1 1
8 0 1 1 1
02/14
LASCO-C2 Level 0.5
SIRW October 23-24, 2003 Pattern Recognition Techniques Applied to Solar Image DataPattern Recognition Techniques Applied to Solar Image Data
1998.06.02
0 1 2 3
0 0 0 0 0
1 1 1 1 1
2 1 1 1 1
... 1 1 1 1
7 1 1 1 1
8 1 1 1 1
0 1 2 3
0 1 1 1 1
1 8 8 8 8
2 8 8 8 8
... 8 8 8 8
7 8 8 8 8
8 8 8 8 8
0 1 2 3
0 1 1 1 1
1 8 18 18 18
2 8 18 18 18
... 8 18 18 18
7 8 18 18 18
8 8 18 18 18
0 1 2 3
0 1 1 1 1
1 18 8 8 8
2 18 8 8 8
... 18 8 8 8
7 18 8 8 8
8 18 8 8 8
0 1 2 3
0 0 0 0 0
1 0 1 1 1
2 0 1 1 1
... 0 1 1 1
7 0 1 1 1
8 0 1 1 1
03/14
LASCO-C2 Level 0.5
SIRW October 23-24, 2003 Pattern Recognition Techniques Applied to Solar Image DataPattern Recognition Techniques Applied to Solar Image Data
1998.06.02
0 1 2 3
0 0 0 0 0
1 1 1 1 1
2 1 1 1 1
... 1 1 1 1
7 1 1 1 1
8 1 1 1 1
0 1 2 3
0 1 1 1 1
1 8 8 8 8
2 8 8 8 8
... 8 8 8 8
7 8 8 8 8
8 8 8 8 8
0 1 2 3
0 1 1 1 1
1 8 18 18 18
2 8 18 18 18
... 8 18 18 18
7 8 18 18 18
8 8 18 18 18
0 1 2 3
0 1 1 1 1
1 18 8 8 8
2 18 8 8 8
... 18 8 8 8
7 18 8 8 8
8 18 8 8 8
0 1 2 3
0 0 0 0 0
1 0 1 1 1
2 0 1 1 1
... 0 1 1 1
7 0 1 1 1
8 0 1 1 1
04/14
LASCO-C2 Level 0.5
SIRW October 23-24, 2003 Pattern Recognition Techniques Applied to Solar Image DataPattern Recognition Techniques Applied to Solar Image Data
1998.06.02
0 1 2 3
0 0 0 0 0
1 1 1 1 1
2 1 1 1 1
... 1 1 1 1
7 1 1 1 1
8 1 1 1 1
0 1 2 3
0 1 1 1 1
1 8 8 8 8
2 8 8 8 8
... 8 8 8 8
7 8 8 8 8
8 8 8 8 8
0 1 2 3
0 1 1 1 1
1 8 18 18 18
2 8 18 18 18
... 8 18 18 18
7 8 18 18 18
8 8 18 18 18
0 1 2 3
0 1 1 1 1
1 18 8 8 8
2 18 8 8 8
... 18 8 8 8
7 18 8 8 8
8 18 8 8 8
0 1 2 3
0 0 0 0 0
1 0 1 1 1
2 0 1 1 1
... 0 1 1 1
7 0 1 1 1
8 0 1 1 1
05/14
LASCO-C2 Level 0.5
SIRW October 23-24, 2003 Pattern Recognition Techniques Applied to Solar Image DataPattern Recognition Techniques Applied to Solar Image Data
1998.06.02
0 1 2 3
0 0 0 0 0
1 1 1 1 1
2 1 1 1 1
... 1 1 1 1
7 1 1 1 1
8 1 1 1 1
0 1 2 3
0 1 1 1 1
1 8 8 8 8
2 8 8 8 8
... 8 8 8 8
7 8 8 8 8
8 8 8 8 8
0 1 2 3
0 1 1 1 1
1 8 18 18 18
2 8 18 18 18
... 8 18 18 18
7 8 18 18 18
8 8 18 18 18
0 1 2 3
0 1 1 1 1
1 18 8 8 8
2 18 8 8 8
... 18 8 8 8
7 18 8 8 8
8 18 8 8 8
0 1 2 3
0 0 0 0 0
1 0 1 1 1
2 0 1 1 1
... 0 1 1 1
7 0 1 1 1
8 0 1 1 1
06/14
LASCO-C2 Level 0.5
SIRW October 23-24, 2003 Pattern Recognition Techniques Applied to Solar Image DataPattern Recognition Techniques Applied to Solar Image Data
1998.06.02
0 1 2 3
0 0 0 0 0
1 1 1 1 1
2 1 1 1 1
... 1 1 1 1
7 1 1 1 1
8 1 1 1 1
0 1 2 3
0 1 1 1 1
1 8 8 8 8
2 8 8 8 8
... 8 8 8 8
7 8 8 8 8
8 8 8 8 8
0 1 2 3
0 1 1 1 1
1 8 18 18 18
2 8 18 18 18
... 8 18 18 18
7 8 18 18 18
8 8 18 18 18
0 1 2 3
0 1 1 1 1
1 18 8 8 8
2 18 8 8 8
... 18 8 8 8
7 18 8 8 8
8 18 8 8 8
0 1 2 3
0 0 0 0 0
1 0 1 1 1
2 0 1 1 1
... 0 1 1 1
7 0 1 1 1
8 0 1 1 1
07/14
LASCO-C2 Level 0.5
SIRW October 23-24, 2003 Pattern Recognition Techniques Applied to Solar Image DataPattern Recognition Techniques Applied to Solar Image Data
1998.06.02
0 1 2 3
0 0 0 0 0
1 1 1 1 1
2 1 1 1 1
... 1 1 1 1
7 1 1 1 1
8 1 1 1 1
0 1 2 3
0 1 1 1 1
1 8 8 8 8
2 8 8 8 8
... 8 8 8 8
7 8 8 8 8
8 8 8 8 8
0 1 2 3
0 1 1 1 1
1 8 18 18 18
2 8 18 18 18
... 8 18 18 18
7 8 18 18 18
8 8 18 18 18
0 1 2 3
0 1 1 1 1
1 18 8 8 8
2 18 8 8 8
... 18 8 8 8
7 18 8 8 8
8 18 8 8 8
0 1 2 3
0 0 0 0 0
1 0 1 1 1
2 0 1 1 1
... 0 1 1 1
7 0 1 1 1
8 0 1 1 1
08/14
LASCO-C2 Level 0.5
SIRW October 23-24, 2003 Pattern Recognition Techniques Applied to Solar Image DataPattern Recognition Techniques Applied to Solar Image Data
1998.06.02
0 1 2 3
0 0 0 0 0
1 1 1 1 1
2 1 1 1 1
... 1 1 1 1
7 1 1 1 1
8 1 1 1 1
0 1 2 3
0 1 1 1 1
1 8 8 8 8
2 8 8 8 8
... 8 8 8 8
7 8 8 8 8
8 8 8 8 8
0 1 2 3
0 1 1 1 1
1 8 18 18 18
2 8 18 18 18
... 8 18 18 18
7 8 18 18 18
8 8 18 18 18
0 1 2 3
0 1 1 1 1
1 18 8 8 8
2 18 8 8 8
... 18 8 8 8
7 18 8 8 8
8 18 8 8 8
0 1 2 3
0 0 0 0 0
1 0 1 1 1
2 0 1 1 1
... 0 1 1 1
7 0 1 1 1
8 0 1 1 1
09/14
LASCO-C2 Level 0.5
SIRW October 23-24, 2003 Pattern Recognition Techniques Applied to Solar Image DataPattern Recognition Techniques Applied to Solar Image Data
1998.06.02
0 1 2 3
0 0 0 0 0
1 1 1 1 1
2 1 1 1 1
... 1 1 1 1
7 1 1 1 1
8 1 1 1 1
0 1 2 3
0 1 1 1 1
1 8 8 8 8
2 8 8 8 8
... 8 8 8 8
7 8 8 8 8
8 8 8 8 8
0 1 2 3
0 1 1 1 1
1 8 18 18 18
2 8 18 18 18
... 8 18 18 18
7 8 18 18 18
8 8 18 18 18
0 1 2 3
0 1 1 1 1
1 18 8 8 8
2 18 8 8 8
... 18 8 8 8
7 18 8 8 8
8 18 8 8 8
0 1 2 3
0 0 0 0 0
1 0 1 1 1
2 0 1 1 1
... 0 1 1 1
7 0 1 1 1
8 0 1 1 1
10/14
LASCO-C2 Level 0.5
SIRW October 23-24, 2003 Pattern Recognition Techniques Applied to Solar Image DataPattern Recognition Techniques Applied to Solar Image Data
1998.06.02
0 1 2 3
0 0 0 0 0
1 1 1 1 1
2 1 1 1 1
... 1 1 1 1
7 1 1 1 1
8 1 1 1 1
0 1 2 3
0 1 1 1 1
1 8 8 8 8
2 8 8 8 8
... 8 8 8 8
7 8 8 8 8
8 8 8 8 8
0 1 2 3
0 1 1 1 1
1 8 18 18 18
2 8 18 18 18
... 8 18 18 18
7 8 18 18 18
8 8 18 18 18
0 1 2 3
0 1 1 1 1
1 18 8 8 8
2 18 8 8 8
... 18 8 8 8
7 18 8 8 8
8 18 8 8 8
0 1 2 3
0 0 0 0 0
1 0 1 1 1
2 0 1 1 1
... 0 1 1 1
7 0 1 1 1
8 0 1 1 1
11/14
LASCO-C2 Level 0.5
SIRW October 23-24, 2003 Pattern Recognition Techniques Applied to Solar Image DataPattern Recognition Techniques Applied to Solar Image Data
1998.06.02
0 1 2 3
0 0 0 0 0
1 1 1 1 1
2 1 1 1 1
... 1 1 1 1
7 1 1 1 1
8 1 1 1 1
0 1 2 3
0 1 1 1 1
1 8 8 8 8
2 8 8 8 8
... 8 8 8 8
7 8 8 8 8
8 8 8 8 8
0 1 2 3
0 1 1 1 1
1 8 18 18 18
2 8 18 18 18
... 8 18 18 18
7 8 18 18 18
8 8 18 18 18
0 1 2 3
0 1 1 1 1
1 18 8 8 8
2 18 8 8 8
... 18 8 8 8
7 18 8 8 8
8 18 8 8 8
0 1 2 3
0 0 0 0 0
1 0 1 1 1
2 0 1 1 1
... 0 1 1 1
7 0 1 1 1
8 0 1 1 1
12/14
LASCO-C2 Level 0.5
SIRW October 23-24, 2003 Pattern Recognition Techniques Applied to Solar Image DataPattern Recognition Techniques Applied to Solar Image Data
1998.06.02
0 1 2 3
0 0 0 0 0
1 1 1 1 1
2 1 1 1 1
... 1 1 1 1
7 1 1 1 1
8 1 1 1 1
0 1 2 3
0 1 1 1 1
1 8 8 8 8
2 8 8 8 8
... 8 8 8 8
7 8 8 8 8
8 8 8 8 8
0 1 2 3
0 1 1 1 1
1 8 18 18 18
2 8 18 18 18
... 8 18 18 18
7 8 18 18 18
8 8 18 18 18
0 1 2 3
0 1 1 1 1
1 18 8 8 8
2 18 8 8 8
... 18 8 8 8
7 18 8 8 8
8 18 8 8 8
0 1 2 3
0 0 0 0 0
1 0 1 1 1
2 0 1 1 1
... 0 1 1 1
7 0 1 1 1
8 0 1 1 1
13/14
LASCO-C2 Level 0.5
SIRW October 23-24, 2003 Pattern Recognition Techniques Applied to Solar Image DataPattern Recognition Techniques Applied to Solar Image Data
1998.06.02
0 1 2 3
0 0 0 0 0
1 1 1 1 1
2 1 1 1 1
... 1 1 1 1
7 1 1 1 1
8 1 1 1 1
0 1 2 3
0 1 1 1 1
1 8 8 8 8
2 8 8 8 8
... 8 8 8 8
7 8 8 8 8
8 8 8 8 8
0 1 2 3
0 1 1 1 1
1 8 18 18 18
2 8 18 18 18
... 8 18 18 18
7 8 18 18 18
8 8 18 18 18
0 1 2 3
0 1 1 1 1
1 18 8 8 8
2 18 8 8 8
... 18 8 8 8
7 18 8 8 8
8 18 8 8 8
0 1 2 3
0 0 0 0 0
1 0 1 1 1
2 0 1 1 1
... 0 1 1 1
7 0 1 1 1
8 0 1 1 1
14/14
LASCO-C2 Level 0.5
SIRW October 23-24, 2003 Pattern Recognition Techniques Applied to Solar Image DataPattern Recognition Techniques Applied to Solar Image Data
LASCO-C22002.08.12 00:06 – 2002.08.14 02:06
Original
0 1 2 3
0 1 1 1 1
1 0 5 5 5
2 0 5 5 5
... 0 5 5 5
7 0 5 5 5
8 0 5 5 5
0 1 2 3
0 0 0 0 0
1 1 1 1 1
2 1 1 1 1
... 1 1 1 1
7 1 1 1 1
8 1 1 1 1
0 1 2 3
0 0 0 0 0
1 0 3 3 3
2 0 3 3 3
... 0 0 0 0
7 0 0 0 0
8 0 0 0 0
1 2
3 4
SIRW October 23-24, 2003 Pattern Recognition Techniques Applied to Solar Image DataPattern Recognition Techniques Applied to Solar Image Data
1 2 3 4
SIRW October 23-24, 2003 Pattern Recognition Techniques Applied to Solar Image DataPattern Recognition Techniques Applied to Solar Image Data
Highlights
• Typical coronal images show coexistent structures exhibiting high and low intensities, i.e., a wide dynamic range.
Method´s property of being highly localized (depending upon the value of N, i.e., size of the mother wavelet, relative to the image size) allows to treat them on the same ground and without affecting each other.
Radial distance from the border of the occulter (in pixels)
As shown with the examples, the SC2003 technique is suitable for the selective enhancement of specific spatial scales composing any 2D
image.
An example showing the application of the SC2003 technique to a polar representation of a LASCO C2 image
SIRW October 23-24, 2003 Pattern Recognition Techniques Applied to Solar Image DataPattern Recognition Techniques Applied to Solar Image Data
Towards automatic tracking of dynamical events
The temporal evolution of dynamical events can be seen in a single 2D image by stacking one image on top of the other and obtaining the intensity profile along the time axis i) at a given position angle for all radial distances (Heigth - Time maps), or ii) at a given radial distance for all position angles (Position angle - Time maps), so that the SC2003 technique can be applied. Two examples using LASCO-C2 data follows .
OriginalProcessed
(Continuum component removed)
Posit
ion
An
gle
Rad
ial d
ista
nce (
pxls
)
Time (hours since January 8, 2001 at 00:06 UT)0 25 50 75 100 125 Time (hours since January 8, 2001 at 00:06 UT)
0 25 50 75 100 125
0
50
100
150
90
270
110
130
150
170
190
210
230
250
SIRW October 23-24, 2003 Pattern Recognition Techniques Applied to Solar Image DataPattern Recognition Techniques Applied to Solar Image Data
270
300
330
0
30
60
90
Posit
ion
An
gle
Rad
ial d
ista
nce (
pxls
) 150
100
50
00 25 50 75 100 125
0 25 50 75 100 125 Time (hours since January 21, 2001 at 00:06 UT)
0 25 50 75 100 125 0 25 50 75 100 125
OriginalProcessed
(Continuum component removed)CMEs
Streamers
Top: Position Angle - Time map Bottom: Corresponding Height-Time Map for a radial cut at P=98° -Example 1- and
P=285° -Example 2- (solid white line in Position Angle - Time map). The dashed white line depicts the border of the occulter.
SIRW October 23-24, 2003 Pattern Recognition Techniques Applied to Solar Image DataPattern Recognition Techniques Applied to Solar Image Data
Future Prospects...Future Prospects...
Time-lapse sequences of LASCO-C2, and C3 show a continuous outflow resembling the flow of the slow solar wind. However, the small inhomogeneities forming the flow cannot be distinguished from noise when observing individual images.
This side effect can be used for good to enhance the inhomogeneities forming the outward flow. As the object to be characterize is a flow, static images will not reveal anything unless the dynamic is in the image itself (e.g., Carrington maps, or Height-Time maps).
If there is no small-scale inhomogeneities moving outward the Heigth-Time image will exhibit just white noise. Otherwise, the background will exhibit a preferential direction (noise correlated in time).
Under way... The SC2003 technique to enhance such inhomogeneities to help quantify the slow solar wind speed...
Without proper noise removal, the technique developed also enhances the noise
Note the inclination of the pattern !!!!!
Time Time
Rad
ial d
ista
nce (
pxls
)
An
gu
lar
dis
tan
ce
SIRW October 23-24, 2003 Pattern Recognition Techniques Applied to Solar Image DataPattern Recognition Techniques Applied to Solar Image Data
END
SIRW October 23-24, 2003 Pattern Recognition Techniques Applied to Solar Image DataPattern Recognition Techniques Applied to Solar Image Data
1/4
SIRW October 23-24, 2003 Pattern Recognition Techniques Applied to Solar Image DataPattern Recognition Techniques Applied to Solar Image Data
2/4
SIRW October 23-24, 2003 Pattern Recognition Techniques Applied to Solar Image DataPattern Recognition Techniques Applied to Solar Image Data
3/4
SIRW October 23-24, 2003 Pattern Recognition Techniques Applied to Solar Image DataPattern Recognition Techniques Applied to Solar Image Data
4/4
SIRW October 23-24, 2003 Pattern Recognition Techniques Applied to Solar Image DataPattern Recognition Techniques Applied to Solar Image Data
- Splines : piecewise polynomials
- Spline degree n : each segment is a polynomial of degree n (n+1 coef needed).Additional smoothness constraint: continuity of the spline and
derivatives until order n-1.
- B splines: basic atoms by which splines are constructed
- B3 minimum curvature property.
Why BWhy B33 splines as mother wavelets? splines as mother wavelets?
2DB3-spline
SIRW October 23-24, 2003 Pattern Recognition Techniques Applied to Solar Image DataPattern Recognition Techniques Applied to Solar Image Data
SIRW October 23-24, 2003 Pattern Recognition Techniques Applied to Solar Image DataPattern Recognition Techniques Applied to Solar Image Data
(1) The arrival of photons, and their expression by electron counts on CCD detectors may be modeled by a Poisson distribution. If the noise in data I(x,y) is Poisson, the Anscombe transformation acts as if the data arose from Gaussian white noise model.
Determination of the Noise
(3) Noise Progression in wavelet scales: by simulating an image containing Gaussian noise with a standard deviation equal to 1, and taking the same WT applied to the original image to this sintetic image. is the standard deviation of each wavelet scale.
j
(2) Calculation of local standard deviation: For a fixed pixel position, say ( k,h ), the local standard deviation is calculated by taking its N x N neighbouring pixels given by the cartesian product [ k,k+N ] x [ h,h+N ] and computing their standard deviation. This value is stored in an array at its corresponding position, i.e., ( k,h ). The operation is extended to cover all pixels, the resulting array being the local standard deviation
),( yxI
(4) Example:
Original image
Original image + white noise (gaussian)
comparable to that of the original signal
After filtering by application of the multiresolution
approach
SIRW October 23-24, 2003 Pattern Recognition Techniques Applied to Solar Image DataPattern Recognition Techniques Applied to Solar Image Data
0 1 2 3
0 1 1 1 1
1 5 0 0 0
2 5 0 0 0
3 5 0 0 0
4 5 0 0 0
5 5 0 0 0
0 1 2 3
0 1 1 1 1
1 5 0 0 0
2 5 0 0 0
... 5 0 0 0
14 5 0 0 0
15 5 0 0 0
SIRW October 23-24, 2003 Pattern Recognition Techniques Applied to Solar Image DataPattern Recognition Techniques Applied to Solar Image Data
0 1 2 3
0 1 1 1 1
1 0 0 0 0
2 0 0 0 0
3 5 5 5 5
... 5 5 5 5
15 5 5 5 5
Original
SIRW October 23-24, 2003 Pattern Recognition Techniques Applied to Solar Image DataPattern Recognition Techniques Applied to Solar Image Data
Original
0 1 2 3
0 1 1 1 1
1 5 0 0 0
2 5 5 5 5
3 5 5 5 5
4 5 5 5 5
5 5 5 5 5
SIRW October 23-24, 2003 Pattern Recognition Techniques Applied to Solar Image DataPattern Recognition Techniques Applied to Solar Image Data
SIRW October 23-24, 2003 Pattern Recognition Techniques Applied to Solar Image DataPattern Recognition Techniques Applied to Solar Image Data
FTs (Fourier Transforms):
- Transforms data from time to frequency domain
- Functions as superpositions of sin and cos
Non-localized
WFTs (Windowed Fourier Transforms):
- Signal is chopped into sections for separate analysis
- Windowing via weight functions
- Gives information both in time and frequency domain