After the first ISOCAM results on relatively strong sources (ISO A&A 1996, 315, special issue) it has become clear that the Camera on ISO could achieve more sensitive observations by integrating longer, particularly on "empty" high galactic latitude fields. Already sources with fluxes larger than the mJy level can routinely be observed which are 1000 times fainter than previous IRAS 12 surveys and with a resolution which is 10 times higher. Reduction of the data requires that effort has to be spent on dealing with CAM peculiarities which show up at these levels. The now "old" mid-infrared detectors on ISO, which were designed ten years ago, suffer from long-term memory effects that have to be dealt with, before the noise can be integrated down. We will show that this is possible because the camera is ultimately highly linear when stabilised, and the so-called transient effects are highly reproducible under the same conditions, and because the zodiacal background, which dominates the high galactic latitude sky in the long wavelength (LW) channel, appears to be smooth on the 10 arcsecond scale. As many important projects with ISOCAM (several hundred hours from both guaranteed and open time proposals) aim at looking for faint extragalactic objects, it is worth developing and demonstrating reliable, independent data reduction techniques. Our approach is complementary to the method presented by Starck et al. (1998) and Aussel et al. (1997, 1998). We will here present how the raw data are processed to remove the main instrumental effects (Sect. 2), and how the triple beam-switch reduced data are projected onto the sky and sources are extracted from the resulting sky map (Sect. 3). At each step, for illustrative purposes, we take examples from the HDF public domain ISOCAM data (Rowan-Robinson et al. 1997). All of our algorithms were developed on Dec-Alpha stations using specific routines written in the IDL language. We then apply the whole ISOCAM data reduction method to the HDF specific case and give a list of the faintest mid-infrared sources that have been observed to date (Sect. 4). Finally we briefly discuss the advantages and drawbacks of this method and the various ways it could be improved.
© European Southern Observatory (ESO) 1999
Online publication: February 22, 1999