Separating detection and catalog production

TL;DR

This paper proposes a modular approach to astronomical data processing, separating detection from catalog production, to enhance flexibility and scientific creativity in large survey data analysis.

Contribution

It introduces a novel modular system with separate software for detection and catalog creation, enabling more flexible and creative analysis pipelines.

Findings

Non-parametric detection at very low surface brightness levels

Modular system improves pipeline flexibility

Empowers scientific experimentation

Abstract

In the coming era of massive surveys (e.g. LSST, SKA), the role of the database designers and the algorithms they choose to adopt becomes the decisive factor in scientific progress. Systems that allow/encourage users/scientists to be more creative with the reduction/analysis algorithms can greatly enhance scientific productivity. The separation/modularity of the detection processes and catalog production is one proposal for achieving `Reduction/analysis algorithms for large databases and vice versa' (a key theme for the 26th ADASS). With the new noise-based detection paradigm, non-parametric detection is now possible for astronomical objects to very low surface brightness limits. In our implementation, one software (NoiseChisel) is in charge of detection and another (MakeCatalog) is in charge of catalog production. This modularity has many advantages for pipeline developers, and more importantly, it empowers scientific curiosity and creativity.