Template matching method for the analysis of interstellar cloud structure

TL;DR

This paper introduces a fast, robust template-matching method for analyzing filamentary and other structures in interstellar medium maps, demonstrating its effectiveness compared to existing techniques like RHT.

Contribution

The paper presents a novel template-matching algorithm for map analysis that is more robust and versatile than previous methods, capable of identifying various structures at different scales.

Findings

TM method is comparable to RHT in performance

TM is more robust to input parameter variations

TM can identify diverse structures at multiple scales

Abstract

The structure of interstellar medium can be characterised at large scales in terms of its global statistics (e.g. power spectra) and at small scales by the properties of individual cores. Interest has been increasing in structures at intermediate scales, resulting in a number of methods being developed for the analysis of filamentary structures. We describe the application of the generic template-matching (TM) method to the analysis of maps. Our aim is to show that it provides a fast and still relatively robust way to identify elongated structures or other image features. We present the implementation of a TM algorithm for map analysis. The results are compared against rolling Hough transform (RHT), one of the methods previously used to identify filamentary structures. We illustrate the method by applying it to Herschel surface brightness data. The performance of the TM method is found to be comparable to that of RHT but TM appears to be more robust regarding the input parameters, for example, those related to the selected spatial scales. Small modifications of TM enable one to target structures at different size and intensity levels. In addition to elongated features, we demonstrate the possibility of using TM to also identify other types of structures. The TM method is a viable tool for data quality control, exploratory data analysis, and even quantitative analysis of structures in image data.