Algorithm for star extraction based on self-adaptive background prediction

Through analysis of the characteristics of target star point and the background in star map, a new algorithm for star extraction based on self-adaptive background prediction is proposed. To predict the background accurately and give attention to the algorithm speed, the regional maximum background prediction is only used at the edge of the background and the fixed weight background prediction is used in other parts. The residual star map is segmented and the target area of the star point is extracted. The threshold value can be obtained self-adaptively. The precise location of the target star is calculated by the centroid algorithm. The experiments show that the performance of the proposed algorithm is superior to the other existing star extracting algorithms. As an effective star extracting algorithm, it can avoid many false alarms caused by strong edge and can predict the background accurately.