KMS of Xinjiang Astronomical Observatory, CAS
| 射电天文干涉图像的快速高保真重建 | |
| Alternative Title | Fast High-fidelity Reconstruction of Radio Astronomical Interferometric Images |
张利
| |
| Subtype | 博士 |
| Thesis Advisor | 刘祥 ; 张明 |
| 2016 | |
| Degree Grantor | 中国科学院大学 |
| Place of Conferral | 北京 |
| Degree Discipline | 天体物理 |
| Keyword | 射电综合成图 快速反卷积 高保真重建 Clean算法 |
| Abstract | 最近几年，射电干涉测量已经有了巨大的发展。新的望远镜有更高的灵敏度，并且能测得更弱的辐射和更多细节能被探测到。这是令人兴奋的。然而，这也导致了更大的技术挑战。为了充分揭示数据中信息，它会变成大图像。对于大图像的重建，存在的算法的性能是非常有限的。这个论文项目的目的是为了发展快速高保真的重建算法,让它有能力处理大图像。快速的收敛是为了解决 大图像重建问题。这是通过引进解析式模型计算实现的。高保真重建是为了提高重建性能。这是通过引进自适应循环机制来实现的。我们提出了新的算法，这些算法都是基于CLEAN重建框架实现的。 我们发现自适应尺度像素分解算法中计算开销最大的部分，并引进了解析的方法来估计潜在的真实流量，也改进了估计初始值的方法。这些使得计算量大大减少（当成图质量是可比较的时候）。现代的干涉阵的点扩展函数的旁瓣已经显著减少，这允许我们做如此的近似来减少计算量。使得它有能力处理更大的图像。这将有利于揭示数据的信息。测试显示我们的算法的计算量很弱地依赖图像的尺度，这是对于处理大图像是非常有利的。测试也显示它能很好地恢复各种尺度的信息。 我们通过引进自适应循环增益技术来提高模型的保真度。在这个论文中，我们介绍两个自适应循环增益机制。一个是基于信噪比的自适应循环增益，另一个是基于形状依赖的自适应循环增益。这等价于在重建过程中引入正则化。测试显示自适应循环增益能够显著提高模型的保真度。这种提高源于对数据内部特征的考虑。 这些新的算法已经应用到模拟观测的EVLA数据，它们已经显示更优的性能。我们正在实现这些算法在CASA 软件中，以便其他用户可以使用。 |
| Other Abstract | In recent years, there has been a rapid development in radio interferometry. The new telescopes have much higher sensitivity. Weaker emission and more details can be detected. This is exciting. However, this also leads to bigger technique challenges. The higher sensitivity can help to detect weaker object, which leads to larger image information. For the reconstruction of large images, the existing reconstruction algorithms have limited performances. The goal of this thesis is to develop fast high-fidelity reconstruction algorithms with a reasonable memory footprint, to make it be able to deal with large images. A fast convergence is to solve the reconstruction problem of large images. It is implemented by introducing analytical computation of model. High fidelity reconstruction is to improve the imaging quality. It is implemented by introducing adaptive loop gain schemes. Our new algorithms have been developed in CLEAN reconstruction framework。 We identify the most expensive step in the Asp-Clean algorithm, and introduce an analytical method to model extended emission, and a modified method for estimating the initial values used for the fitting procedure, which significantly reduces the computational load while keeping the imaging performance comparable to the original algorithm. The PSF sidelobe levels of modern arrays are significantly reduced, allowing us to make approximations to reduce the computing load, which in turn allows for the deconvolution of larger images in reasonable timescales. Tests shown that the runtime scales of the new algorithm is only very weakly dependent on the total image size, which is beneficial to large images, and different scales can be effectively recovered. We improve the model fidelity by introducing adaptive loop gain techniques. We introduce two adaptive loop gain schemes in this dissertation. One is adaptive loop gain based on the signal to noise ratio, and another is a shape-dependent adaptive loop gain. These are equivalent to introduce regularization into reconstruction process. Tests shown that adaptive loop gain scheme can significantly improve the model fidelity. The improvements derive from the consideration of internal characteristics of data. Our new algorithms have been applied to simulated EVLA observations and have shown superior performance. The new algorithms are being implemented in CASA package and will be available for all users in the near future. |
| Language | 中文 |
| Document Type | 学位论文 |
| Identifier | http://ir.xao.ac.cn/handle/45760611-7/1726 |
| Collection | 研究生学位论文 射电天文研究室 |
| Affiliation | 中国科学院新疆天文台 |
| Recommended Citation GB/T 7714 | 张利. 射电天文干涉图像的快速高保真重建[D]. 北京. 中国科学院大学,2016. |
| Files in This Item: | ||||||
| File Name/Size | DocType | Version | Access | License | ||
| 张利-学位论文.pdf（11691KB） | 学位论文 | 暂不开放 | CC BY-NC-SA | Application Full Text | ||
Items in the repository are protected by copyright, with all rights reserved, unless otherwise indicated.
Edit Comment