A Multiscale Technique for Automatically Detecting and Tracking CMEs in Coronagraph Data

Studying coronal mass ejections (CMEs) in coronagraph data can be challenging due to their diffuse structure and transient nature, and user-specific biases may be introduced through visual inspection of the images. The large amounts of data available from the SOHO, STEREO, and future Solar Orbiter missions, also makes manual cataloguing of CMEs tedious, and so a robust method of detection and analysis is required. This has led to the development of automated CME detection and cataloguing packages such as CACTus, SEEDS and ARTEMIS. However, the main drawbacks of these catalogues are: the CACTus method of detection fails to resolve CME acceleration profiles; the CACTus and SEEDS running-difference images suffer from spatiotemporal crosstalk; and the SEEDS and ARTEMIS detections are limited to only the LASCO/C2 field-of-view. Recently, the benefits of multiscale filtering of coronagraph data have been demonstrated in an effort to overcome current cataloguing issues. A multiscale decomposition can be applied to individual images in order to enhance the structure of CMEs whilst removing noise and small-scale features like stars. Here we present the development of a new, automated, multiscale, CME detection & tracking technique. It works by first separating the dynamic CME signal from the background corona and then characterising CME structure via a multiscale edge-detection algorithm. The detections are then chained through time to determine the CME kinematics and morphological changes as it propagates across the plane-of-sky. We demonstrate its application to a sample of LASCO data and prove its efficacy in detecting and tracking CMEs. This technique is being applied to the complete LASCO dataset, and it is planned to further develop it for implementation on the SECCHI/COR dataset in the near future.