[Previous] | [Session 54] | [Next]
D. Esch (Department of Statistics, Harvard University)
To make the best of Chandra data of extended sources one wants to ``remove" the point-spread function. Yet it is difficult to properly represent complex extended images such as supernova remnants with wisps, knots, and shells. Hence we are using a very general ``non-parametric" method called a Markov Random Field. The procedure uses a Markov-chain Monte Carlo (MCMC) technique, which has the advantage of providing estimates of the uncertainty in the smoothed images. The procedure works by conditioning on an observed image file, assuming that these counts are Poisson observations contaminated with Poisson background counts, and further assuming that adjacent pixels in the ``true image" have Gaussian differences in intensity on the log scale. This Gaussian density can be tuned, by fixing the variance parameter appropriately, to allow for greater or lesser smoothing in the intensities of the image. A major advantage of the MCMC fitting technique is that the deconvolution of the PSF and the estimation of the true image intensities happen simultaneously in the fitting algorithm, thus reducing ``black box" error that can occur when procedures to accomplish these ends are applied sequentially. Example images from the Chandra X-ray Telescope will be presented, and compared, using different smoothness parameters in the model.
If you would like more information about this abstract, please follow the link to http://www.fas.harvard.edu/~vandyk/astrostat.html. This link was provided by the author. When you follow it, you will leave the Web site for this meeting; to return, you should use the Back comand on your browser.
The author(s) of this abstract have provided an email address for comments about the abstract: esch@stat.harvard.edu