On the trail of a comet's tail: A particle tracking algorithm for comet 67P/Churyumov-Gerasimenko

TL;DR

This paper presents a novel particle tracking algorithm for analyzing debris motion in images from the Rosetta mission to comet 67P, enabling detailed velocity and acceleration measurements of particles in the comet's coma.

Contribution

We developed a new algorithm that detects and tracks particles in image sequences, allowing for the analysis of their dynamics and origins in the comet's inner coma.

Findings

Identified 2268 particle tracks in a sample sequence.

Approximately 52% of tracks are likely genuine based on manual and simulated data analysis.

First results on particle velocity, acceleration, and size distributions consistent with previous studies.

Abstract

Context. During the post-perihelion phase of the European Space Agency's Rosetta mission to comet 67P, the Optical, Spectroscopic, and Infrared Remote Imaging System on board the spacecraft took numerous image sequences of the near-nucleus coma, with many showing the motion of individual pieces of debris ejected from active surface areas into space. Aims. We aim to track the motion of individual particles in these image sequences and derive their projected velocities and accelerations. This should help us to constrain their point of origin on the surface, understand the forces that influence their dynamics in the inner coma, and predict whether they will fall back to the surface or escape to interplanetary space. Methods. We have developed an algorithm that tracks the motion of particles appearing as point sources in image sequences. Our algorithm employs a point source detection software to locate the particles and then exploits the image sequences' pair-nature to reconstruct the particle tracks and derive the projected velocities and accelerations. We also constrained the particle size from their brightness. Results. Our algorithm identified 2268 tracks in a sample image sequence. Manual inspection not only found that 1187 (~52%) of them are likely genuine, but in combination with runs on simulated data it also revealed a simple criterion related to the completeness of a track to single out a large subset of the genuine tracks without the need for manual intervention. A tentative analysis of a small (n = 89) group of particles exemplifies how our data can be used, and provides first results on the particles' velocity, acceleration, and radius distributions, which agree with previous work.