Weekly Report 17Sep2010
Imaging the Albedo with Pixon
The following show Pixon images for simulated data of a circular source with albedo. Each image shows the contours for the Pixon reconstruction overlaid on a map of the original data. Below each image are the profiles for the data and the Pixon fit to the profile, along with the C statistic for each detector. For each image the following parameters were used:
- Time: 2-September-2002 00:07:36 - 00:08:20
- Energy Interval: 24.0 keV - 51.0 keV
- Detectors: All Detectors
- Image Size: 64 X 64
- Pixel Size: .5 X .5
Above each image is the total counts in the final image. Below each image is the source size data as calculated using the 3 sigma source size program written by Rick Pernak.
Reconstructed Pixon Images for a point source with albedo located at heliocentric X = 301 arcsec and Y = 0.0
The total counts in the final image below is: ~10e4 counts
The 3 sigma parameters for the reconstructed image assuming an elliptical source are:
Source Flux: 32.0174 Source X: 301.042 Source Y: -0.0517814 Source FWHM: 4.71759 Eccentricity: 0.253152 Source Angle: 90.0000
The following plot shows the profile for each detector compared with the fitted profile along with the C statistic for each detector.
The total counts in the final image below is: ~10e5 counts
The 3 sigma parameters for the reconstructed image assuming an elliptical source are:
Source Flux: 310.592 Source X: 301.157 Source Y: -0.344853 Source FWHM: 3.94633 Eccentricity: 0.0599250 Source Angle: 107.000
The following plot shows the profile for each detector compared with the fitted profile along with the C statistic for each detector.
The total counts in the final image below is: ~10e6
The 3 sigma parameters for the reconstructed image assuming an elliptical source are:
Source Flux: 1473.15 Source X: 301.232 Source Y: -0.429101 Source FWHM: 4.42663 Eccentricity: 0.260343 Source Angle: 167.000
The following plot shows the profile for each detector compared with the fitted profile along with the C statistic for each detector.
The plot below shows the flux profile of each image along a line parallel to the Heliocentric X axis. Each solid line shows the profile for an image with different total counts. The dashed line shows the profile of the original data map scaled to match the peak of the trace with the highest flux.
Reconstructed Pixon Images for a point source with albedo located at heliocentric X = 929 arcsec and Y = 0.0
The following image shows the original data map with contours for a reconstructed Pixon image with total counts of ~10e4 near the limb.
The 3 sigma parameters for the reconstructed image assuming an elliptical source are:
Source Flux: 31.2434 Source X: 930.370 Source Y: -0.653541 Source FWHM: 3.15973 Eccentricity: 0.623439 Source Angle: 197.000
The following plot shows the profile for each detector compared with the fitted profile along with the C statistic for each detector.
The following image shows the original data map with contours for a reconstructed Pixon image with total counts of ~10e5 near the limb.
The 3 sigma parameters for the reconstructed image assuming an elliptical source are:
Source Flux: 311.739 Source X: 930.703 Source Y: -0.720957 Source FWHM: 3.15171 Eccentricity: 0.642290 Source Angle: 0.00000
The following plot shows the profile for each detector compared with the fitted profile along with the C statistic for each detector.
The following image shows the original data map with contours for a reconstructed Pixon image with total counts of ~10e6 near the limb.
The 3 sigma parameters for the reconstructed image assuming an elliptical source are:
Source Flux: 1532.89 Source X: 930.723 Source Y: -0.440351 Source FWHM: 2.73080 Eccentricity: 0.650604 Source Angle: 183.000
The following plot shows the profile for each detector compared with the fitted profile along with the C statistic for each detector.
The plot below shows the flux profile of each image along a line parallel to the Heliocentric X axis. Each solid line shows the profile for an image with different total counts. The dashed line shows the profile of the original data map scaled to match the peak of the trace with the highest flux.
Conclusions
- In the case of both the images near the limb at 929 arcsecs and closer to disk center at 301 arcsecs the C statistic increases as the total counts in the image increases. In the coarser detectors the increase is greater. Richard Schwartz suggested that this is because with more counts differences between the fitted profile and the actual profile are no longer within statistical error but can be treated as actual differences between the fit and the measured counts/bin thereby driving up the C statistic. Does this mean that when we have images with higher total counts the detectors with coarser grids cannot be used to create an image?
- The FWHM and eccentricity from the 3 sigma fit show the foreshortening of the circular source. For the case of the circular source located at a Heliocentric X value of 301 arcsecs, the FWHM is between ~4.7 arcsecs and ~4 arcsecs depending on the total counts. The eccentricity of the ellipse is between ~.05 and ~.2. This means that the source is close to circular with a FWHM of around 4 arcsecs. This is close to the original data of a circular source with a FWHM of approximately 5 arcsecs. For the case of the circular source near the limb at a Heliocentric X value of 929 arcescs, the 3 sigma fit of an ellipse to the source yields a FWHM between ~2.7 arcsecs and ~3.2 arcsecs with an eccentricity of ~.6. This means that the source appears compressed and elliptical when near the limb as would be expected.
- The flux profiles of the image at a Heliocentric X value of 301 arcsecs are symmetric around the peak of the profile. They have a FWHM of around ~4 arcsecs. When the profile from the original data map is scaled to the peak of the profile with the highest flux they match each other, although the original data map has a smoother profile. When looking at the profiles for the images near the limb at 929 arcsecs it appears that any foreshortening only shows up as the total counts in the image increase. At a total counts of ~10e4 the profile looks symmetric. As the total counts go up to ~10e5 and ~10e6 asymmetry starts to show up in the profile. It spreads preferentially towards the disk center displaying the foreshortening effect. Comparing the profile of the image with the highest total counts to the profile for the original data map shows that the original map has a narrower peak and and the shoulder along the Heliocentric X direction towards disk center drops off rapidly. It is possible that the peak becomes too narrow and the drop off is too rapid to be imaged by Pixon. If the count rate is high enough Pixon may be able to demonstrate foreshortening but cannot image it correctly.
Populating the new RHESSI Website
Continued with populating the new RHESSI Website. Links were added back to documents on the old site. Some of these documents will be left as is, and some will be moved to the new site. The following pages were update:
Finishing RHESSI Inventory
RHESSI inventory was finished. Items which were not on our original property list were accounted for and the inventory list was returned to the 671 administration.
Goals for next week
- Use other algorithms with new simulated data files to create images. Check to see if albedo can be imaged and whether the dimensions of the images agree with the original data files using the 3 sigma fit algorithm.
- Continue populating the new RHESSI Website with content
- Check the prepserver logs to track problems with xrt_prep on hesperia
- Continue with documentation for new website editors