Presentation #332.01 in the session The Sun and Solar System II.
As the discovery rate of solar system objects (SSOs) rises, questions regarding their properties continue to emerge and researchers face challenges of identifying observations of both new and known SSOs among survey data. Covering three quarters of the sky, the National Science Foundation’s NOIRLab Source Catalog (NSC, Nidever et al 2018, 2021) combines sources detected in exposures from multiple separate surveys and instruments. The first (255,454 exposures) and second (412,116 exposures) data releases of the NSC span 7 and 9 years of observation respectively, offering sufficient spatial and temporal coverage for the presence of SSO ephemerides in the data. To differentiate SSOs from more distant bodies in the NSC, all measurements are classified as representing a “stationary” (distant sources) or “moving” object (nearby SSOs), largely through the iterative use of the Density-Based Spatial Clustering of Applications with Noise (DBSCAN, Ester et al. 1996) available through scikit-learn (Pedregosa et al. 2011). The first application of DBSCAN largely identifies objects external to the solar system, treated as “stationary”. DBSCAN clusters detections spaced <5 arcseconds apart, leaving all other measurements as unclustered outliers. Each cluster’s proper motion is calculated to determine the object’s likelihood of being “stationary” or a slow-moving SSO. Two methods were developed to identify individual moving objects within the remaining unclustered data by forming roughly linear “tracklets”, each a group of 3+ measurements representing a unique object. The Computationally Automated NSC tracklet Finder (CANFind, Fasbender & Nidever, 2021) applies a second iteration of DBSCAN, clustering measurements spaced 0.5-10.8 arcseconds apart to form tracklets. The Tough Transform (Fasbender & Nidever, in prep.), an adaptation of the Hough Transform (Radon, 1917), associates detections that are even more spatially separated from one another by having sources “vote” for configurations most likely to represent SSO motion. Tracklets detected in NSC DR1 and DR2 using CANFind and the Tough Transform are validated by their calculated proper motion and fit to solar system orbits, creating SSO candidates to be further analyzed.