In order to facilitate usage of optical data in space climate studies, we have developed an automated algorithm to quantify the complexity of auroral structures as they appear in ground-based all-sky images. The image analysis is based on a computationally determined “arciness” value, which describes how arc like the auroral structures in the image are. With this new automatic method we have analyzed the type of aurora in about 1 million images of green aurora (λ = 557.7nm) captured at five camera stations in Finnish and Swedish Lapland in 1996–2007. We found that highly arc like structures can be observed in any time sector and their portion of the auroral structures varies much less than the fraction of more complex forms. The diurnal distribution of arciness is in agreement with an earlier study with high arc occurrence rate in the evening hours and steadily decreasing toward the late morning hours. The evolution of less arc-like auroral structures is more dependent on the level of geomagnetic activity and solar cycle than the occurrence of arcs. The median arciness is higher during the years close to the solar minimum than during the rest of the solar cycle. Unlike earlier proposed, the occurrence rate of both arcs and more complex auroral structures increases toward the solar maximum and decreases toward the solar minimum. The cyclic behavior of auroral structures seen in our data is much more systematic and clear than previously reported visual studies suggest. The continuous arciness index describing the complexity of auroral structures can improve our understanding on auroral morphology beyond the few commonly accepted structure classes, such as arcs, patches, and omega bands. Arciness can further be used to study the relationship of auroral structures at different complexity levels and magnetospheric dynamics.