Detection and spatial analysis of fairy circles

In this paper, fairy circles are identified in aerial image data using image analysis approaches. The phenomena of fairy circles attracts research to investigate the causes of these abnormal patterns of plant growth. We detect (segment) fairy circles in available image data, and measure properties in terms of size, shape, and spatial distribution. Different segmentation techniques are applied on high-resolution image data taken with an unmanned aerial vehicle (UAV). Low-resolution images from freely available Apple maps are also used to discuss spatial distribution and the size of fairy circles. Experimental results suggest that methods used in this paper can efficiently segment fairy circles in image data. Moreover, a 2D Gaussian-like distribution can be found in the spatial distribution of the fairy circles in Namibia and Australia with respect to the size of circles, but with a noticeable difference in distribution parameters when comparing both locations of fairy circles.