The acronym PISA stands for Position, Intensity and Shape Analysis, and is the group name for a package of routines that deal with the location and parameterisation of objects on an image frame.
The core of this package is the routine PISAFIND which performs image analysis on a 2-dimensional data frame. The program searches the data array for objects that have a minimum number of connected pixels above a given threshold and extracts the image parameters (position, intensity, shape) for each object. The image parameters can be determined using thresholding techniques or an analytical stellar profile can be used to fit the objects. In crowded regions deblending of overlapping sources can be performed.
PISAFIND is based on the APM IMAGES routine. The algorithms are the same as in IMAGES but the interface has been adapted to Starlink, so that for instance the input data is now expected to be an NDF (see [1] which also describes other features of the Starlink Software Environment). The APM object finding and analysis package IMAGES was originally written by Mike Irwin at the University of Cambridge to analyse output from the Automated Photographic Measuring system. The background to the APM image analysis package is given in ??[2].
In addition to its object detection capability PISA can also perform simple object classification. The classification is performed by the simple thresholding of an additionally generated ‘peakedness’ measure, or by the multivariate analysis of a set of intensity corrected variables.