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

Automatic candidate detections

The use of an automatic procedure for source extraction from real images is mandatory when very large mosaic images are involved (about 64 million pixels) and the source properties need to be compared in different colours.

We refer the reader to Bertin and Arnouts (1996) for a detailed description of SExtractor software. Here we wish to focus on the determination of the best extraction parameters, based on the image quality and noise characteristic of our image data sample.

SExtractor gives the possibility of filtering images with a kernel so that the noise is decreased and the S/N increased: this kernel is chosen in such a way that it reproduces the PSF of the stars in the field. For the OACDF extraction, our choice has been for a Gaussian kernel, whose FWHM properly matches the seeing on the images. Detection is performed on the filtered image. The detection threshold was determined as the best compromise between limiting magnitude and percentage of spurious detection.



Juan Alcala
2002-02-05