| i2gaussfit | stsdas.analysis.fitting | i2gaussfit |
i2gaussfit -- Iterative 2-dimensional Gaussian fit to noisy image data.
i2gaussfit input output
This is a script task that fits a 2-dimensional Gaussian to one image. In noisy data, the standard n2gaussfit task usually must be used more than once, fitting each Gaussian parameter separately until convergence is met. In this kind of data, wrong results can be generated when simultaneously fitting all parameters when the initial guess is not very near the solution. This task calls n2gaussfit successively in a loop, fitting each parameter in the following order:
1. Fit amplitude and background, keep center and fwhm fixed. 2. Fit center, keep everything else fixed. 3. Fit fwhm, keep everything else fixed. 4. Go back to (1), repeat the loop 'niter' times. 5. Fit all parameters simultaneously.The amoeba fitting method is used. In the fifth step, coefficient errors can be estimated by using the errors parameter. Results at each individual step can be displayed at the terminal using the verbose parameter. The task uses a temporary table in the current directory for storing intermediate results.
The remaining parameters needed for the fit must be defined using the n2gaussfit parameter file, and the tgausspars pset. In particular, starting conditions can be input from the pset or from a previously obtained table row, as usual in n2gaussfit. The ellipticity parameters ellip and theta are read from the pset or table, but are kept fixed during the fitting loop. (See the n2gaussfit help page for details.)
1. Fit a 2-dimensional Gaussian to a section of the image test and store the fit results in the file testfit.db in the user's home directory:
fi> i2gaussfit test[100:500,256:300] home$testfit.db
This task was written by I.Busko
