Increasing the Spectral Resolution of Broadband Astronomical Imaging

最新一代的望远镜和仪器正在打开一扇了解可变和瞬息万变的宇宙的新窗口。即将到来的时间域调查将收集大部分夜空的数千个观测纪元。但这些数据提供了超越时间序列研究的机会，该项目将致力于实现更高分辨率的形状测量的承诺，并改进对星系属性的推断。初步工作已经显示出未来成功的希望。应用数学、计算机科学和天文学的学生将联手应对真实数据和微妙影响的挑战。多时代的图像和视频分析可以说是当今数据科学中增长最快的话题。在这项研究中创造的形象和方法是激励和训练学生的理想选择。这些视觉表现也是公众出游的有效工具。超越瞬变和可变宇宙的科学首先是通过重叠、重复曝光的过采样来实现的，这可以产生更高的空间分辨率和形状测量，从而改进弱透镜、宇宙学和星系形态研究。不那么明显的是，在一个宽通带内引入多个有效子带将显著改善对星系性质的推断和对光度红移的估计。这项研究将致力于实现这些机会所需的新方法，办法是：(1)评估这些想法在实际情况中的局限性，包括大型天文观测望远镜的测光红移估计；(2)开发最佳、有效的重建方法；(3)实施生产质量编码，从多历元天文观测中制作超分辨率超彩色图像。建立在稳健统计基础上的方法已经在单色环境中取得了成功，但多色将需要对大气中的差色折射(DCR)进行详细的建模。这将需要一种新的方法来利用由于DCR引起的源方向上与颜色相关的细微变化。创建的部分教育材料将以Python格式公开提供Jupyter笔记本。科学数据的彩色渲染将被开发为实时Jupyter笔记本，并在本网站上提供给所有人，只需使用浏览器即可查看和修改。该奖项反映了NSF的法定使命，并通过使用基金会的智力优势和更广泛的影响审查标准进行评估，被认为值得支持。

The latest-generation telescopes and instruments are opening a new window on the variable and transient universe. Forthcoming time-domain surveys will collect thousands of observational epochs for most of the night sky. But these data provide opportunities beyond the study of time series, and this project will work towards realizing the promise of higher resolution shape measurement and improving the inference of galaxy properties. Preliminary work has already shown the promise of future success. Students in applied mathematics, computer science and astronomy will join forces to tackle the challenges of real data and subtle effects. Multi-epoch image and video analysis is arguably the fastest growing topic in data science today. The images and methods created in this study are ideal for inspiring and training students. The derived visual representations are also an efficient vehicle for public outreach.The science beyond the transient and variable cosmos comes firstly by oversampling in overlapping, repeated exposures, which can produce higher spatial resolution and shape measurements, thus improving weak lensing, cosmology and galaxy morphology studies. Less obviously, introducing multiple effective sub-bands within a broad passband will significantly improve the inference of galaxy properties and the estimation of photometric redshifts. This study will work towards the new methods needed to realize these opportunities, by (1) evaluating the limitations of these ideas in real situations, including photometric redshift estimation for the Large Synoptic Survey Telescope (LSST); (2) developing optimal, efficient methods for reconstruction; and (3) implementing production-quality code to make super-resolution hyper-color images from multi-epoch astronomical observations. The approach of building on robust statistics has already been successful in a monochrome setting, but multi-color will need detailed modeling of differential chromatic refraction (DCR) in the atmosphere. This will require a new method to leverage the color-dependent subtle shifts in source directions due to DCR. Part of the created educational material will be publicly available Jupyter notebooks in Python. Colorful renderings of scientific data will be developed as live Jupyter notebooks and made available on the SciServer for all to view and to modify, using just a browser.This award reflects NSF's statutory mission and has been deemed worthy of support through evaluation using the Foundation's intellectual merit and broader impacts review criteria.