Optimizing ISOCAM data processing using spatial redundancy

Several instrumental effects of the Long Wavelength channel of ISOCAM, the camera on board the Infrared Space Observatory, degrade the processed images. We present new data-processing techniques that correct these defects, taking advantage of the fact that a position in the sky has been observed by several pixels at different times. We use this redundant information (1) to correct the long-term variation of the detector response, (2) to correct memory effects alter glitches and point sources, and (3) to refine the deglitching process. As an example we have applied our processing to the gamma-ragi burst observation GRB 970402. Our new data-processing techniques allow the detection of faint extended emission with contrast smaller than 1% of the zodiacal background. The data reduction corrects instrumental effects to the point where the noise in the final map is dominated by the: readout and the photon noises. All raster ISOCAM observations call benefit from the data processing described here. This includes mapping of solar system extended objects (comet dust trails), nearby clouds and star forming regions! images from diffuse emission in the Galactic plane and external galaxies. These techniques could also be applied to other raster type observations (g.g. ISOPHOT).