The K-Vector ND and its Application to Building a Non-Dimensional Star Identification Catalog
Conference Paper
Overview
Research
Identity
Additional Document Info
Other
View All
Overview
abstract
A multi-dimensional orthogonal range-searching algorithm, the multi-dimensional k-vector (K-Vector ND), is presented. The algorithm is analyzed and found to have an execution time that is independent of the size of the database, for well-distributed data sets. Numerical tests are performed to determine the performance advantage as compared to a Quad-Tree for the two-dimensional data set. Results range from break-even to a factor of 14, depending on the database size. The K-Vector ND is then applied to the problem of building a non-dimensional star-identification database that contains all visible star triples. The performance of the K-Vector ND algorithm in that task is then compared to a simple nested loop, and found to range from break-even to a factor of 200, depending on the size of the database.