星图识别是星光辅助导航中关键的一步。目前大部分星图目标检测算法均针对模拟星空图进行研究,且往往采用单一算法,算法的实际可行性和效果值得怀疑。针对实际拍摄的星空图,先后利用维纳滤波、数学形态学滤波、软阂值小波去噪和阈值分割等算法进行星图目标检测。处理结果表明了算法的有效性。
Star map identification is a key step in CNS. The existing star detection methods which almost use single algorithm aim at simulation image,the practical feasibility and effect need to be proved. The paper advancs a new detecting algorithm by using Wiener filter,mathematical morphology lilting, soft threshold wavelet d enoising and threshold segmenting algorithm orderly, based on the practical star image. The processing results show the algorithm is effectual.