射电星表与多波段数据自动化证认方法研究

不同波段星表交叉证认是天文数据融合及天文学研究中颇为棘手的问题。如果被证认的源具有延展结构，则更是如此。例如，在射电波段观测的射电星系具有核与喷流等结构。这些结构的相互关系不易确认，也难以在其他波段星表中找到对应体。以往这类证认工作只能由射电天文学家人工完成。这种工作方式无法应对ALMA、FAST、SKA等新一代观测设备带来的海量数据。因而，本项目提出一种自动化方法来证认射电天体的各个部分及其在其他波段的对应体。我们通过数学模型来描述可能相关的射电源，使用贝叶斯假设推断方法对模型进行概率计算以检查它们的相关性，并找到被证认源在其他波段的对应体。这种自动证认方法提供了多波段延展天体证认的可行途径，可应用于未来大规模巡天观测数据的交叉证认和数据融合工作。项目组计划将本方法应用于SWIRE红外星表和ATLAS射电星表的重叠区域上来检验新方法的有效性，并与此前天文学家人工证认的结果进行对比分析。

Catalogs crossmatching among different wavelengths is a common but very difficult problem in astronomy, especially when objects have extended spatial structures like jet, lobe, etc. In radio observation, an object can have several detections corresponding to various components, and it is hard to specify the relation between them, i.e. to cross-identify among different wavelengths. Traditionally, these associations are annotated by hand by expert radio astronomers. This however is not practical nor scalable to the massive datasets resulting from the next generation advanced telescopes, e.g. ALMA, FAST, SKA. We propose to develop an automated method for finding the correct association between radio detections. Based on geometrical model, we can include related sources and find the possibility of their combination by Bayesian hypothesis testing. To test the effectiveness of our method, we would crossmatch the SWIRE infrared catalog with the ATLAS radio catalog. Then to verify the accuracy of our new method we compare our results to the result of manual cross identification by radio astronomers. We expect this method will provide a fast, accurate and scalable solution for crossmatching and combining catalogs of different wavelengths.