The standard ISOCAM reduction method consists of the following steps:
Deglitching. On any observation, numerous energetic particle impacts leave traces on one or several pixels of the detector. Most of these impacts do not affect the detector response on long time scale and are seen as instantaneous flux step (fast glitches). In Fig. 2, a typical flux history of one pixel as a function of time is shown where many fast glitches are apparent. Various techniques are used to suppress these emission spikes (Starck 1999).
Dark image subtraction. The response of the LW detector of ISOCAM is not zero when the detector does not receive any photons coming from the sky. This dark image must be subtracted from the data. Temporal variation of the dark image has been observed (Starck et al. 1999) and modeled by Biviano et al. (1998) using a polynomial function of time.
Short term transient.
It has been known since pre-launch measurements (Perault et al. 1994) that the ISOCAM detector is affected
by a transient effect, i.e. its response to a given flux step is not instantaneous.
Recently a new model of the response, based on the physic of Si:Ga arrays (Fouks & Schubert 1995),
has been developed by Coulais & Abergel (2000).
The uncertainty in the corrected flux with this new model is
1% but
raises to about 10% for point sources, likely due to charge coupling effects between
neighboring pixels.
This is not a strong problem here since we are mostly interested in diffuse emission.
Flat Field. Generally, each pixel of a detector does not have exactly the same response to a given flux. This spatial variation of the detector response (the flat field in the following) must be taken into account. The most straightforward way to compute an image of the flat field is to average all readouts of a data cube, and normalize the mean to 1 (Starck et al. 1999). This technique supposes that all camera pixels have seen the same average flux along the observation. It is a reasonable approximation in raster mode where the camera is moving on a region of the sky much larger than the field of view of the camera, which does not contain any systematic gradient.
Building the sky image. After dividing the data cube by the flat field, the final standard data reduction step is to project each readout onto the sky to build the sky image. The LW channel of ISOCAM is affected by a field of view distortion problem that must be taken into account in this operation (Starck et al. 1999), especially to properly add emission from point sources and small scale structures.
The sky image of the GRB1 observation obtained with the standard data processing described in this section is shown in Fig. 8A. It is affected by many instrumental problems that limit the full sensitivity of the instrument. In the next section we describe the open problems for the imaging of faint extended emission.
Copyright The European Southern Observatory (ESO)