Using active contours for automated tracking of UV and EUV solar flare ribbons

QB Astronomy;QC Physics

Gill, Colin D. ; Fletcher, Lyndsay

University:University of Glasgow

Department:School of Physics and Astronomy

关键词: solar flares, flare ribbons, EUV ribbons, UV ribbons, TRACE, image processing, active contours, flare, footpoints;   

Others  :  http://theses.gla.ac.uk/975/1/2009gillmsc.pdf

来源: University of Glasgow

【 摘 要 】

Solar flare UV and EUV images show elongated bright ``ribbons'' that move over time.If these ribbons are assumed to locate the footpoints of magnetic field lines reconnecting in the corona then it is clear that studying their evolution can provide an important insight into the reconnection process.We propose an image processing method based on active contours (commonly referred to as ``snakes''), for tracking UV and EUV flare ribbons in images from the transition region and coronal explorer (TRACE).Our method aims to provide an efficient, accurate and automatic tool to aid in the study of flare ribbon evolution in large datasets.Chapter 1 provides an introduction to the Sun and solar activity, with a more focussed section on solar flares where the mechanism for the creation of flare ribbons is discussed.We also outline the motivation for solar physics research as a whole and more specifically the motivations behind this project.In this chapter we introduce the TRACE satellite as the source the images used in this project, with a summary of its hardware and the UV and EUV channels which the images used in this project are captured in.Chapter 2 introduces some basics of image processing, such as applying spatial filters.We also look at the different approaches to image segmentation including a more in depth study of active contours.The role of image processing in solar physics and the driving forces for image processing development in solar physics are summarised.The final section of this chapter presents a review of previous methods used for the tracking of solar flare ribbons, including manual, semi-automatic and fully automatic methods.The first part of Chapter 3 details the pre-processing steps applied to the TRACE data before its use by our algorithm.The second part of this chapter introduces our algorithm, with a general overview and detailed discussion of the constituent parts.We also show results from initial tests carried out using a simulated test image, and demonstrate how different parameters of the algorithm can affect its result.Chapter 4 shows results obtained from using our algorithm on TRACE flare images.Some modifications to the algorithm were deemed necessary after applying it to only a small number of flare images, the initial part of this chapter covers the reasons for the modifications and the modifications themselves.The remainder of the chapter presents results of the algorithm applied to a number image sequences from different flares.The results are presented and discussed for each flare separately, with one flare being used as an example of how the parameters of the algorithm can be adapted to suit different flares and images.Chapter 5 discusses to what extent the aim of the project has been achieved and presents a summary of the problems encountered in applying our algorithm to flare images.This chapter finishes with a look at some ideas for future work, both for our algorithm specifically, and for general efforts at flare ribbon tracking.