Star Pattern Identification Algorithm using Linear Fractional Transformation of Triangle Similarity

摘要

A new algorithm to identify star patterns using a linear fractional transformation which maps similar triangles to a point in a unit disk on the complex plane. The algorithm uses identification of star-triangles similar to the triangles made from the input points and the intersection of two star-ID permutations sharing the same stars. The algorithm is coded in MATLAB to develop a star pattern identification system named TRIASTER. The experiment results show the effectiveness of the algorithm. It accepts a rough sketch of starry sky as the input data when the radius of tolerance circle on the unit disk is 0.3 and the number of input points is 6 or 7.