Application of Radial Symmetry for Caldera Detection

For geoscientific applications, automatic shape-based feature detection offers a fast and non-subjective means of identifying geological structures within data. Most previously published examples of circular feature detection for geoscientific applications aimed to identify impact craters from optical or topographic data. Various techniques used include the Hough transform, template matching, and machine learning. This paper presents a circular feature detection technique based on the fast radial symmetry transform, which is applied to volcanic caldera detection using topographic data. The detection method combines the fast radial symmetry transform with high-level interpretation. Our system identifies areas with both elevated and depressed circular landforms using the radial symmetry measures and detects the size of the summit crater using an energy minimising spline. An experiment using data from Papua New Guinea demonstrated that the technique can accommodate calderas with variations such as incomplete rims. The technique used is capable of rapidly processing large datasets as an initial screening tool to assist a manual identification process.