基于限定搜索高斯曲面拟合的星点质心提取算法  

Star centroid extraction algorithm based on limited search Guassian surface fitting

作　　者：赵峥[1,2] 丁飞[1,2] 李泽群 周启航 章华涛[3,4] 张登银[1,2] ZHAO Zheng;DING Fei;LI Zequn;ZHOU Qihang;ZHANG Huatao;ZHANG Dengyin(Jiangsu Key Laboratory of Broadband Wireless Communication and Internet of Things,Nanjing University of Posts and Telecommunications,Nanjing 210003,China;Intelligent Internet of Things Application Technology Research Institute,Nanjing University of Posts and Telecommunications,Nanjing 210003,China;Nanjing Institute of Astronomical Optics&Technology,National Astronomical Observatories,Chinese Academy of Sciences,Nanjing 210042,China;CAS Key Laboratory of Astronomical Optics&Technology,Nanjing 210042,China)

机构地区：[1]南京邮电大学江苏省宽带无线通信和物联网重点实验室,江苏南京210003 [2]南京邮电大学智慧物联网应用技术研究院,江苏南京210003 [3]中国科学院国家天文台南京天文光学技术研究所,江苏南京210042 [4]中国科学院天文光学技术重点实验室,江苏南京210042

出　　处：《中国测试》2023年第12期54-59,共6页China Measurement & Test

基　　金：国家自然科学基金(61872423);南京邮电大学科研创新基金NUPTSF(NY220028)。

摘　　要：星点光斑图像目标识别技术对于空间监视装备和系统至关重要,由于星点图像中光斑成像的特点以及不同背景下图像噪声的干扰,目标光斑的自动识别和定位精度受限。高斯曲面拟合法与密度质心法是使用较为广泛的星点光斑目标定位算法,通过理论分析和实验表明传统方法对星点质心的定位存在一定误差。该文提出一种基于限定区域搜索的星点质心提取算法(LASACE),该算法首先通过中值滤波算法与最大熵阈值分割法对星点图像进行预处理,然后使用高斯曲面拟合算法获得星点光斑的像素分布。最后以星点中心建立目标搜索区域,采用距离强加权质心法提取出目标星点光斑的质心位置。实验结果表明,相较于密度质心法、高斯曲面拟合法和加权质心法,该文所提算法在降低星点光斑质心定位误差的同时提高了抗噪性,能够满足空间卫星的识别需求。The technology of star point image spot target recognition is crucial for space monitoring equipment and systems.Due to the characteristics of spot imaging in star point images and the interference of image noise in different backgrounds,the accuracy of automatic recognition and positioning of target spots is limited.Gaussian surface fitting method and density centroid method are widely used star spot target positioning algorithms.Theoretical analysis and experiments have shown that traditional methods have certain errors in locating star spot targets.A star centroid extraction algorithm based on limited area search(LASACE)algorithm is proposed.The algorithm first preprocesses the star image through median filtering algorithm and maximum entropy threshold segmentation method,and then uses Gaussian surface fitting algorithm to obtain the target center and pixel distribution of the star spot.Finally,the target search area is established using star point light spots,and the position of the target light spot is obtained using the distance strong weighted centroid method.The experimental results show that compared to the density centroid method,Gaussian surface fitting method,and weighted centroid method,the proposed algorithm can improve the positioning and recognition accuracy of star points by more than one pixel.The algorithm improves noise resistance while reducing star positioning errors,and can meet the recognition needs of space satellites.

关 键 词：星点图像 光斑中心 高斯曲面拟合 质心法 

分 类 号：TP391.41[自动化与计算机技术—计算机应用技术] V448.22[自动化与计算机技术—计算机科学与技术] TB9[航空宇航科学与技术—飞行器设计]