Automatic classification of auroral images in substorm studies

Millions of auroral images are captured every year by ground-based imagers. Even though the auroral appearance or “type” yields relevant information about the physical processes in the ionosphere and the magnetosphere, qualitative descriptions of auroras are typically used. Modern methods including those widely used in computer vision research can, however, make it possible to use objective and quantitative measures in analysing auroral appearance. We are currently developing techniques for automated auroral image analysis. In order to numerically compare auroral objects, we can either describe individual auroral shapes — such as arcs — or use statistical appearance models (texture). We demonstrate how one can use Fourier Descriptors to compare shapes extracted from auroral images. Also, using a recently developed texture analysis technique, we show how texture measurements can be used in classifying auroral type in a timeseries.