Date: Monday 07-November-2022
Time: 11:15-12:15
Location: Zoom (contact for link)
Speaker: Mohammad Akhlaghi
Title: Carving out the low surface brightness astronomical signal
Abstract: Astronomical instrumentation has greatly advanced over the last 40 years: with digital detectors, space telescopes and +8m class ground-based telescopes for example. However, the signal-based detection paradigm is still the dominant method of low-level data analysis (for example from Petrosian or Kron in the 1970s, mostly through the implementation in SExtractor from the mid-1990s): detection, segmentation and measurements or catalog production. In this talk, after reviewing the major systematic biases regarding astronomical object detection and segmentation that is inherent to the signal-based paradigm, a fundamentally different "noise based" detection paradigm will be introduced for detecting signal that may have extremely low signal-to-noise ratios. With thresholds that are below the Sky value, and non-parametric expansion into noise, it is successfully able to detect very diffuse and irregularly shaped signal in noise (e.g., comets, nebulae, or galaxies) and improve the sky-level estimation by one order of magnitude see [1]. The software implementation of this method is called NoiseChisel (part of GNU Astronomy Utilities, or Gnuastro [2]). To define sub-structure over the detected signal, another software (Segment, also within Gnuastro) is in charge of Segmentation, finding true "clumps" over the NoiseChisel detections and using those to define "objects", see [3]. The talk will conclude with the application of NoiseChisel in several research projects and impact on science outputs.
[1] https://doi.org/10.1088/0067-0049/220/1/1 [2] https://www.gnu.org/software/gnuastro [3] https://arxiv.org/abs/1909.11230
For further information contact PALS coordinator Quanzhi Ye at qye@umd.edu.
SPECIAL ACCOMMODATIONS:Special accommodations for individuals with disabilities can be made by calling (301) 405-3001. It would be appreciated if we are notified at least one week in advance.
