Using clustering for disperse objects fields segmentation in MIRADAS instrument

Abstract

Mid-resolution InfRAreD Astronomical Spectrograph (MIRADAS) is a near-infrared multi-object spectrograph for Gran Telescopio Canarias (GTC). It has 12 deployable Integral Field Units (IFU) based on probe arms with pick-off mirrors, each of which can observe a different user-defined sky object. MIRADAS can work with target sets where their components are spread over such a wide area so that all of them do not fit in the field-of-view of the instrument. Therefore, data sets of that kind require, prior to capturing them, some arrangement that groups its elements in different subsets where the distance between the two most remote elements is inferior to the field-of-view diameter. This field segmentation is achieved using a hierarchical clustering technique. Our method relies on determining mutual nearest-neighbors, which will be merged if they show a given degree of similarity known beforehand. Moreover, we also compute a geometric center for these clusters, information to be delivered to the telescope’s pointing process. This step is formulated as the minimum bounding disk problem, which founds the center of the smallest radius circle enclosing all points of a cluster. Finally, we consider several real science cases and analyze the performance of the proposed solution.