Constellation Recognition on Digital Images

In this paper, we present an optimization-based approach for star constellation recognition. The main components of the proposed procedure are the processing of digital images and the minimization of the error in point cloud matching by parameter optimization. In the optimization phase, the Nelder– Mead algorithm, the simulated annealing algorithm, and an evolutionary algorithm were used. Also, the behavior of these methods is evaluated and compared in the paper based on their average error values for different test images. Results on our test dataset showed that the Nelder–Mead simplex algorithm was performing the best in solving the recognition task.