基于FAST的星图中星的提取方法  

Star Acquisition From Star Image Based on FAST

作　　者：石汉路[1] 付文兴[2] 李凯悦 马杰[1] 田金文[1] 

机构地区：[1]华中科技大学多谱信息处理技术国家重点实验室,武汉430074 [2]北京机电工程研究所,北京100074

出　　处：《计算机与数字工程》2015年第10期1799-1802,共4页Computer & Digital Engineering

基　　金：中央高校基本科研业务费项目(编号:2014GF);航天科技创新基金(编号:CASC04);一院高校联合创新基金(CALT201303)资助

摘　　要：在分析星图中星体目标信号特点的基础上,提出了一种将改进的FAST角点检测算法和自适应阈值分割算法相结合的星图中星的提取方法。首先为FAST算法引入自适应阈值选取方法,以进行不同图像对比度情形下的自动星点检测,并利用候选点邻域内周围像素点的拉普拉斯值,剔除了虚假角点;然后,利用自适应阈值分割方法和连通性分析,在局部邻域内确定星体像素区域;最后,利用质心法完成质心定位。仿真结果表明:这种方法与矢量法相比,效率相当,但性能更好。Based on the characteristic of star target in star image,a new method of star acquisition combining improved FAST corner detection algorithm with auto-adaptive threshold segmentation method is proposed in this paper.Firstly,autoadaptive threshold is introduced into FAST detector to conduct target detection automatically under different contrast and laplacian value of the pixel in the neighborhood of candidate point is used to eliminate some false keypoints.Then,auto-adaptive threshold segmentation algorithm and connectivity are used to determine star pixels in its neighborhood.Finally,centroiding procedure can be conducted to get its position.The simulation results indicate that the proposed method has equivalent efficiency to the vector technique while providing stronger anti-noise ability.

关 键 词：星敏感器 FAST角点检测算法 自适应阈值 

分 类 号：TP391[自动化与计算机技术—计算机应用技术]