Treffer: Efficient Multi-Threaded Data Starting Point Matching Method for Space Target Cataloging

Currently, multi-target survey telescope arrays play an important role in the build-up and maintenance of space object catalog databases, collecting massive observational data without attributing information. However, the matching process of massive observational data poses significant challenges to traditional prediction methods. To address the issues of low matching success rates and prolonged computation times in traditional methods, this paper proposes a multi-threaded data starting point matching method. First, orbital elements from the Space Surveillance and Tracking (SST) database are extracted for two days before and after the observation moment. A set of orbital elements closest to the observation epoch is filtered to form the primary candidate catalog containing the maximum number of objects. A matching error threshold is set. Second, multi-threaded traversal of the primary candidate catalog is performed to calculate observation residuals with the data starting point using an orbit prediction procedure. Orbital elements meeting the triple matching error threshold are selected to form the secondary candidate catalog, which is used in the entire data arc segment-matching calculation. Finally, the root mean square error (RMSE) of observation residuals for the entire data arc segment is computed point by point. The orbital elements satisfying the matching threshold are identified as matching results based on the principle of optimality. Experimental results demonstrate that with a matching error threshold of 1°, the proposed method achieves an average matching success rate of 97.62% for data arc segments with nearly 10,000 passes per day over 8 consecutive days. In the SST database containing an average of 25,720 targets, this method processes an average of 2164 data arc segments per minute, improving matching efficiency by 115 times compared to traditional prediction methods.