The Pixon algorithm used for the reconstruction of images from RHESSI data is an adaptation of the program successfully used to analyze data from Yohkoh/HXT (Metcalf et al. 1996). Unlike Clean, which models the source as a collection of point sources, this algorithm seeks a superposition of circular sources or pixons of different sizes and parabolic profiles that best reproduces the measured modulations from the different detectors. The goal is to construct the image with the fewest degrees of freedom (the fewest pixons) that is consistent with the observations (i.e., the image predicts the modulated count rates with a value of chi-square per degree of freedom acceptably close to one).

This technique is generally thought to provide the most accurate image photometry but is very slow if you are not careful about setting the parameters.  A 128x128 pixel image takes approximately 40 minutes to reconstruct on a 3 GHz processor with the default settings for pixon parameters.  However, by setting the image dimensions to 32x32, and enabling the 'fast' option (by setting o->set,/pixon_variable_metric, or clicking the 'Use Fast Algorithm' button in the GUI),  the image took only 3 minutes to reconstruct. The fast option requires more memory - a 128x128 image reconstruction needs 3-5 GB of memory.

As with the other reconstruction algorithms, the Pixon technique as currently implemented can have difficulties with the most compact sources breaking up when the finest grids are used. This problem can be alleviated with various user-controlled parameters.

The Pixon algorithm uses the ANNSEC coordinate system internally.

Using the Pixon algorithm at the command line:

Select algorithm 

o->set, image_algorithm='pixon'

or 'pix' or 'hsi_pixon'

Object Class 


Extract Object

pixon_obj = o->get(/obj,class='hsi_pixon') 

Extract the object used in the hsi_image object, o

Parameter Prefix 

pixon, e.g.
pixon_rgof = o->get(/pixon_rgof)
pixon_control = o->get(/pixon, /control)
pixon_info = o-> get(/pixon, /info)
pixon_all = o->get(/pixon)

All parameter names specific to this algorithm have this prefix

Parameter Table 

Pixon Object Parameter Table

List and short description of control and info parameters specific to this algorithm