next up previous
Next: Selection of galaxy cluster Up: Catalogue extraction Previous: Completeness magnitude limit of

Spurious detections

The fraction of spurious detections is derived from a simulated image, consisting of a realistic background plus modelled catalogue. A good parameter set for a source detection is the one which produces a small fraction of spurious detections in the last magnitude bin correspondent to the completeness limit. The background of this fully simulated image was obtained from the background interpolated by SExtractor, plus the noise structure of the image. This is obtained from the RMS map given by SExtractor multiplied by a simulated Gaussian noise image with null mean and RMS=1 .

For a given parameter set of SExtractor, the number of spurious detections was determined by the difference between the SEx catalogue and the matched catalogue with the input modelled population. Using the OACDF mosaic of April 1999 in the R band, we find that the number of spurious sources, relative to the total number of detected objects, is less than 2% at the limit of 25.2. In Figure 11 (upper panel) the histogram of detected sources is compared with that of spurious detections. The histogram corresponds to the 3000 modeled objects used to estimate the fraction of spurious detections. Furthermore this last quantity (normalised to number counts in the magnitude bin) is shown in the lower panel of Figure 11.


next up previous
Next: Selection of galaxy cluster Up: Catalogue extraction Previous: Completeness magnitude limit of
Juan Alcala
2002-02-05