Connected groups of pixels are accepted as objects if they have more than a minimum number of pixels. Such groups may be rejected if they contact the edges of the data array.
Threshold estimation is performed using either a percentage data point (i.e. the value for which this percentage of pixels have a lower value) or by using a standard deviation and background value determined by fitting a gaussian to the data histogram.
If the logging system has been initialised using CCDSETUP, then the value specified there will be used. Otherwise, the default is "CCDPACK.LOG". [CCDPACK.LOG]
These may be specified as list of comma separated names, using indirection if required, OR, as a single modification element (of the input image names). The simplest modification element is the asterisk "" which means call each of the output lists the same name as the corresponding input images. So:
INsignifies that all the images in the current directory should be used and the output lists should have the same names.
Other types of modification can also occur, such as:
OUTLIST _objs.datwhich means call the position lists the same as the input images but put "_objs.dat" after the names. Replacement of a specified string with another in the output file names can also be used:
OUTLIST _debias_images.datthis replaces the string "_debias" with "_images.dat" in any of the output names.
If wildcarded names for the input images are used then is it recommended that wildcards are also used for the position list names (the order of input names is not guaranteed).
The output files contain a integer index for each image feature followed by the X and Y centroid (formed using all the intensity information) and finally the mean intensity of pixels in the group. [.DAT]
The histogram used by FINDOBJ when USEPER is FALSE is formed by (if necessary) re-binning until the BINFRAC criterion is met, it is expected that this will always result in a well sampled histogram. The background value is the mode of this histogram and is not refined during the gaussian fitting. The gaussian fitting just estimates the standard deviation of the background and uses a fixed peak value and position (the mode of the histogram) and iterates rejecting bins whose counts fall below 20 percent of the peak value, stopping when either 3 iterations have been performed or the standard deviation does not change by more than one bin width in data values.
FINDOBJ is optimised to determine a reliable detection threshold and is not concerned with the accurate determination of the background value on a frame (as it performs no photometric measurements). For this reason the histogram which it uses to determine the background value is made in such a way that it is usually very well sampled (probably oversampled, for most other purposes). FINDOBJ should not be used in a manner for which it is not suited without understanding how if differs from other more specialized routines.
In all cases the coordinates in position lists are pixel coordinates.
Certain parameters (LOGTO and LOGFILE) have global values. These global values will always take precedence, except when an assignment is made on the command line. Global values may be set and reset using the CCDSETUP and CCDCLEAR commands.