基于背景自适应预测的星点提取算法  

Algorithm for star extraction based on self-adaptive background prediction

作　　者：王洪涛[1] 罗长洲[2] 王渝[1] 王向周[1] 赵述芳 

机构地区：[1]北京理工大学自动化学院,北京100081 [2]北京控制与电子技术研究所,北京100038

出　　处：《光学技术》2009年第3期412-414,418,共4页Optical Technique

摘　　要：在分析星图中的星点目标和背景特性的基础上,提出了一种基于背景自适应预测的星点提取算法。为准确预测背景并兼顾算法的速度,只在背景边缘处使用区域最大值背景预测法,其余部分直接使用固定权值进行预测。分割残差星图,提取各星点的目标区域,阈值采用自适应的取值方式。利用质心算法求取各星点目标的精确位置。实验结果表明,基于背景自适应预测的星点提取算法性能优于现有的星点提取算法,是一种有效的星点提取算法,可以降低大量由于强边缘而引起的虚警,能够更准确地预测背景。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 cen- troid 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.

关 键 词：星敏感器 星点提取 背景自适应预测 

分 类 号：TN911.73[电子电信—通信与信息系统] TP391[电子电信—信息与通信工程]