How to extract the star targets from star image is a challenging problem, particularly when the background is intricate and star target is weak. In this paper, an algorithm based on visual case is proposed to achieve fast extraction of star target. It is necessary to get every pixel in the star area to calculate the central coordinate. The boundary searching method that can get the boundary pixels was developed to get every pixel in star area according to the max and min line number and column number. By establishing visual case, algorithm complexity can be reduced during searching boundary of star target. An improved triangle identification algorithm is proposed to identify star target by storing angular distance and marking the identifying state during searching angular distance library. Experimental results show that the consuming time of visual case method is reduced by 3/4 compared to the boundary searching method, and the identification rate and storage requirements are improved compared with conventional triangle algorithm.
Read full abstract