Characterization of inpaint residuals in interferometric measurements of the epoch of reionization
Characterization of inpaint residuals in interferometric measurements of the epoch of reionization
复制标题
再电离时代干涉测量中修复残差的表征
DOI：
10.1093/mnras/stad441
复制
发表时间：
2023
影响因子：
4.8
通讯作者：
Adams, Tyrone
中科院分区：
文献类型：
--
作者：
Pagano, Michael;Liu, Jing;Liu, Adrian;Kern, Nicholas S;Ewall-Wice, Aaron;Bull, Philip;Pascua, Robert;Ravanbakhsh, Siamak;Abdurashidova, Zara;Adams, Tyrone
为了减少射频干扰对21 cm干涉仪数据分析管道的影响，人们开发了许多内涂技术。在这篇文章中，我们检查了由于修复而引入可见度和功率谱的定性和定量误差。我们对模拟数据和来自再电离阵列氢纪元(HERA)第一阶段上限的真实数据进行了分析。我们还介绍了一种卷积神经网络，它能够修复受RFI破坏的数据。我们在模拟数据上训练我们的网络，并表明我们的网络能够修复真实数据，而不需要重新训练。我们发现，在模型中加入延迟空间中的高波数的技术最适合在窄带RFI上进行修复。我们发现，在我们的基准参数下，离散长椭球序列(DPSS)和离散长椭球序列(CIN)为间歇性RFI提供了最好的性能，而高斯进度回归(GPR)和最小二乘谱分析(LSSA)为较大的RFI间隔提供了最好的性能。然而，我们需要注意的是，这些定性的结论对每种修复技术所选择的超参数是敏感的。我们证明了所有的修复技术都可靠地再现了功率谱中以前景为主的模式。由于修复技术不应该能够再现噪声实现，我们发现最大的误差出现在以噪声为主的延迟模式中。我们表明，随着数据的噪声水平降低，CLEAN和DPS最有能力再现可见度中的精细频率结构。
To mitigate the effects of Radio Frequency Interference (RFI) on the data analysis pipelines of 21 cm interferometric instruments, numerous inpaint techniques have been developed. In this paper, we examine the qualitative and quantitative errors introduced into the visibilities and power spectrum due to inpainting. We perform our analysis on simulated data as well as real data from the Hydrogen Epoch of Reionization Array (HERA) Phase 1 upper limits. We also introduce a convolutional neural network that is capable of inpainting RFI corrupted data. We train our network on simulated data and show that our network is capable of inpainting real data without requiring to be retrained. We find that techniques that incorporate high wavenumbers in delay space in their modelling are best suited for inpainting over narrowband RFI. We show that with our fiducial parameters discrete prolate spheroidal sequences (dpss) andcleanprovide the best performance for intermittent RFI while Gaussian progress regression (gpr) and least squares spectral analysis (lssa) provide the best performance for larger RFI gaps. However, we caution that these qualitative conclusions are sensitive to the chosen hyperparameters of each inpainting technique. We show that all inpainting techniques reliably reproduce foreground dominated modes in the power spectrum. Since the inpainting techniques should not be capable of reproducing noise realizations, we find that the largest errors occur in the noise dominated delay modes. We show that as the noise level of the data comes down,cleananddpssare most capable of reproducing the fine frequency structure in the visibilities.