A Boosting Approach for the Detection of Faint Compact Sources in Wide Field Aperture Synthesis Radio Images

摘要

Several thresholding techniques have been proposed so far in order to perform faint compact source detection in wide field interferometric radio images. Due to their low intensity/noise ratio, some objects can be easily missed by these automatic detection methods. In this paper we present a novel approach to overcome this problem. Our proposal is based on using local features extracted from a bank of filters. These features provide a description of different types of faint source structures. Our approach performs an initial training step in order to automatically learn and select the most salient features, which are then used in a Boosting classifier to perform the detection. The validity of our method is demonstrated using 19 images that compose a 2.5° x 2.5° radio mosaic, obtained with the Giant Metrewave Radio Telescope, centered on the MGRO J2019+37 peak of gamma emission at the Cygnus region. A comparison with two previously published radio catalogues of this region (task SAD of AIPS and SExtractor) is also provided.