摘要
干涉阵列存在的点扩展函数旁瓣使得观测到的射电源出现不同程度的失真,对重建宇宙真实结构图景和理解宇宙起源造成影响。为解决观测中出现的伪影,本文在现有的CLEAN算法的基础上,提出了基于L-BFGS-B局部极小化的自适应尺度CLEAN算法。首先,基于L-BFGS-B局部极小化算法通过最小化目标函数,寻找最优分量,构建自适应尺度模型;其次,通过CASA实现对测试图像的重建,对比目前广泛使用的Hogbom CLEAN算法的重建图像,评估本文算法性能;最后,对反卷积算法在射电天文图像处理领域的发展做出展望。测试结果表明:相比于传统的算法,本文提出的算法能够构建更加精准的天空亮度分布,为天文图像重建提供一种新的方案。
The sidelobes of point-spread function in interferometer arrays make the observed radio sources have varying degrees of distortion,which affects the reconstruction of the real structure of the universe and the understanding of the origin of the universe.In order to solve the artifacts in observation,on the basis of the existing CLEAN algorithm,this paper proposes an adaptive scale CLEAN algorithm based on L-BFGS-B local minimization.Firstly,based on the L-BFGS-B local minimization algorithm,an adaptive scale model is constructed by minimizing the objective function and finding the optimal component;secondly,to evaluate the performance of the proposed algorithm;the test image is reconstructed through CASA,which is compared with the widely used Hogbom CLEAN algorithm;finally,the development of deconvolution algorithm in radio astronomy image processing is prospected.The results show that compared with the traditional algorithm,the algorithm proposed can build a more accurate sky brightness distribution,which provide a new scheme for the reconstruction of astronomical images.
作者
张利
肖一凡
米立功
卢梅
赵庆超
王蓓
刘祥
张明
谢泉
ZHANG Li;XIAO Yifan;MI Ligong;LU Mei;ZHAO Qingchao;WANG Bei;LIU Xiang;ZHANG Ming;XIE Quan(College of Big Data and Information Engineering,Guizhou University,Guiyang 550025,China;Xinjiang Astronomical Observatory,Chinese Academy of Sciences,Urumqi 830011,China)
出处
《贵州大学学报（自然科学版）》
2021年第1期38-44,共7页
Journal of Guizhou University:Natural Sciences
基金
平方公里阵列射电望远镜专项资助项目(2020SKA0110300)
国家重点研发计划资助项目(2018YFA0404602)
国家自然科学基金资助项目(11963003)
贵州省教育厅青年科技人才成长资助项目(黔教合KY字[2018]119)。
