Comparison of automatic methods to detect sunspots in the Coimbra Observatory spectroheliograms

TL;DR

This study compares two automated methods for detecting sunspots in Coimbra Observatory spectroheliograms, aiming to efficiently analyze a large historical solar image dataset.

Contribution

It introduces and evaluates two novel automated techniques for sunspot detection in historical solar images, enhancing data analysis capabilities.

Findings

Both methods achieved high Precision and Recall scores.

The intensity-based approach outperformed the morphological method in accuracy.

The tools showed good agreement with existing solar feature catalogs.

Abstract

The Astronomical Observatory of the University of Coimbra has a huge collection of solar images, acquired daily since 1926. From the beginning, only spectroheliograms in the CaiiK line has been recorded, and since 1989 in the H_alpha line also. Such dataset requires efficient tools to detect and analyze solar activity features. The objective of this work is to create a tool that allows to automatic detect sunspots, umbra, and penumbra, that can be applied to the entire dataset. To achieve this, two different approaches have been developed, one based on mathematical morphology and another based on the intensities of the digital levels of the pixels. Both approaches were applied to a subset of images with features identified visually by an experimented solar observer. The performance of both methods was compared through the metrics Precision, Recall and F-score. Another evaluation was made based on the catalogs from Heliophysics Features Catalog and the SILSO catalogue.