Date:1996-06
Author:E. Bertin and S. Arnouts
Pages:12
Summary:The paper introduces SExtractor, a software designed for the automated analysis of astronomical images. SExtractor is capable of detecting, deblending, measuring, and classifying sources from large images with minimal human intervention. Key features include robust star/galaxy separation using a neural network trained with simulated images, handling of various object shapes and magnitudes, and efficient processing of large survey data. The software has been applied to several photometric galaxy surveys and is particularly suited for analyzing extragalactic surveys. The analysis process involves six steps: background estimation, detection, deblending, filtering, photometry, and star/galaxy separation. The paper details each step, including the background estimation method, detection techniques, deblending algorithms, filtering procedures, photometric methods, and the neural network-based star/galaxy classifier. SExtractor's performance is evaluated through simulations and real images, demonstrating its reliability and efficiency in handling crowded fields and blended objects.