Fuzzy based Approach for Recognition of Solar Coronal Holes in SDO/AIA Images

Abstract:

Solar coronal holes (CHs) play a vital role in maintaining the space weather. It is an important source of the geomagnetic storm. This storm ultimately affects the performance of the Earth’s artificial satellite. Moreover, the geomagnetic storm also affects the GPS navigation system. Alongside it also interrupts the power grid service on the Earth's surface. As such, accurate recognition of solar CHs becomes necessary for the proper prediction and estimation of the geomagnetic storm and its impact on the space weather. These CHs regions on the solar surface are prominent in extra-ultra violet and X-ray wavelength SDO/AIA images. Till the end of the 20th century, scientists preferred manual techniques for the detection of these CHs in the solar disk image. However, with the advancement of image processing technology various automated object segmentation tools have been employed for the extraction of CHs regions from the solar image. However, most of the techniques are having a high computational time as – well – as they have higher mathematical complexity. The presence of a non-discrete CHs boundary, existence of solar-limbs along the boundary of the solar disk also affects the performance of the automated detection techniques. Furthermore, the CHs regions, solar filament channels, and the solar image background are having a similar intensity level, which is a challenging issue for any image segmentation methodologies. To overcome these issues in the way of CHs analysis parameterized online region-based fuzzy active contour model has been proposed for real-time detection of CHs. As the fuzzy-based method has the property to handle the uncertainty and ambiguity associated with the image, and parameterization will force the contour to move towards the CHs boundary by eliminating the redundant regions.