Illustrating the Solar Stormwatch "Storm Front" data processing

This animation illustrates the method by which the classifications produced by the Solar Stormwatch "Storm front" activity are processed to provide an estimate of the coronal mass ejection (CME) front in the STEREO-A Heliospheric Imager 1 (HI1) field of view. It demonstrates how each individual classification is pooled to provide a whole set of classifications, and then these are analsysed statistically to estimate the consensus location of the CME front in the HI field of view. The distribution of the classifications is estimated, and thresholded to find regions where featrues have been identified. A segmentation algorithm is used to select the largest feature, which is assumed to be associated with a CME. This identifies classifications that relate to the CME, and the distriubtion of these points is then used to estimate the elongation of the CME front on a regular grid of position angles, by extracting position angle transects of the full distribution of classifications. The animation ends by showing the resulting estimate of the CME front location laid upon the STEREO-A HI1 differenced image, and shows how the process works for following frames.