射电天文图像的反卷积算法研究  

Study on Deconvolution Algorithm of Radio Astronomical Images

作　　者：张利 徐龙[2] 米立功 马家君 ZHANG Li;XU Long;MI Li-gong;MA Jia-jun(College of Big Data and Information Engineering,Guizhou University,Guiyang 550025;Key Labboratory of Solar Activity,National Astronomical Observatories,Chinese Academy of Sciences,Beijing 100012;School of Physics and Electronics,Qiannan Normal University for Nationalities,Duyun 558000)

机构地区：[1]贵州大学大数据与信息工程学院,贵阳550025 [2]中国科学院太阳活动重点实验室,北京100012 [3]黔南民族师范学院物理与电子科学学院,都匀558000

出　　处：《天文学报》2018年第6期117-124,共8页Acta Astronomica Sinica

基　　金：国家自然科学基金(61605153;61572461;11790305);中国科学院太阳活动重点实验室开放课题(KLSA201805);贵州省科技计划项目(黔科合平台人才[2017]5788号);贵州省科技联合基金(黔科合LH字[2015]7710);天文学本科专业建设项目(2016XBJKX0202);贵州省科技厅自然科学基础研究计划(黔科合LH字[2017]7224);贵州省教育厅青年科技人才成长项目(黔教合KY字[2017]107;[2018]119);贵州大学博士基金(贵大人基合字(2016)61号)项目资助

摘　　要：在射电天文干涉测量中,测量的图像包含设备点扩展函数的影响. CLEAN反卷积算法是移去点扩展函数旁瓣影响的最常用算法.自适应尺度像素分解算法是一种尺度敏感的CLEAN反卷积算法,适合于延展源的重建.然而这种算法是耗时的.实现了一种尺度敏感的反卷积算法.它使用若干高斯函数来逼近潜在的真实天空图像,同时用新的方法估计较小的初始分量.实验表明,算法在获得高质量重建结果的同时,速度提高3倍左右.In the radio astronomical interferometry,a measured image contains the effects of the point spread function of the device.The CLEAN deconvolution algorithm and many variants are the most widely used for removing the sidelobe effects of point spread functions.The adaptive scale pixel decomposition algorithm is a scale-sensitive CLEAN deconvolution algorithm suitable for reconstructing extended sources.The scale adaptation of modeled components is achieved using explicit fitting and active sets,but this algorithm is time consuming.A new scale-sensitive deconvolution algorithm is implemented in this paper.This algorithm uses a fitting technique to find a series of Gaussian functions to approximate a potentially real sky image,but it does not pursue the orthogonalization of the search space.Fitting can separate signals and noise well to achieve a high-quality reconstruction.The non-orthogonalization of the search space requires less computational load.At the same time,a new method is used to estimate the small initial components to reduce the computational load.This algorithm is tested on the simulated data and compared to the other typical CLEAN variants.Experiments show that the algorithm in this paper achieves a three-fold increase in speed while achieving high-quality reconstruction results.

关 键 词：方法 数据分析 技术 图像处理 CLEAN算法 

分 类 号：P164[天文地球—天文学]