Generating airglow images and star identification in AMON ACC data

(The project assignment is presently in preparation)

The aim of the exercise is to give students an opportunity to experiment with ground-based astronomical observations. The project includes and common activity – assignment of star catalogue data to astronomical imagery. Besides being a challenging task, the result would be useful for the analysis of AMON data. The project is an expansion of the example presented in a lab activity (see: Stars subtraction in AMON ACC data).

This project’s main complexity is the association of stars in an image with the catalogue data (so-called astrometric solving). After this identification is solved, one can estimate the centre of rotation (location of the north pole in the image). The rotation is necessary for the compensation of different capture time of G-filter and N-filter images. These images are subtracted to remove the starlight in the image and preserve airglow. However, at this point, one of the images has to be rotated such that the stars on images overlap. The association of stars to pixels should also be used to gain information about the brightness of the observed stars in the image captured by AMON ACC. The measurement of brightness should be plotted as a function of time.

If some of the described tasks present themselves more challenging than expected, and if the issues are well-documented by students, the project might not need to achieve all of the goals stated here.

Complexities of the task

Problems with the association of the catalogue data to the image are likely to be numerous, some of these problems are the following:

  • optical deformation of the image due to the fisheye lens,
  • the unspecified azimuth and zenith orientation of the AMON ACC camera, and in turn, the images captured by it,
  • imprecisions in star positions due to atmospheric effects and quantization errors of the raster image,
  • thresholding and normalization of the brightness values in the image.

There is a possibility that the knowledge that the images are taken 5 or 10 minutes apart (depending on which pair of images/filters is considered; this needs verification) might be useful in the determination of the starting point of the fit procedure.

Students can attempt to do the astrometric solving task also by adapting a third-party solution. Implementation of a custom algorithm is not required. Also, students can use a third-party solution to resolve issues related to optical deformations.

After identifying at least one star in images, the star can be used to determine the centre of rotation in the images, which is the north pole. This star could be Polaris, but the approach should also work with a different star in principle. We can find the centre by fitting the Polaris’ measured positions to an ellipse, and the centre of the ellipse is the centre of rotation.

Example of the estimated centre of rotation from the position of Polaris (Credit: Šimon Mackovjak and his team, IEP SAS)

Then there is the topic of rotation of images in order for star positions to overlap. First, there is a need to know the angle of rotation. Second, there is the consideration that the image is a projection of a sphere.

There should be several ways to determine the angle of rotation. One method is to consider the relative time of the image capture. The images are captured 5 or 10 minutes apart (depending on which pair of images/filters is considered; this needs verification). Another method is to consider the rotation that was needed to fit the star positions to the images. Subtraction of the two angles should indicate the difference in rotation of the two images. Another suitable approach is to measure the angle between the same star positions in two images, considering the fitted centre of rotation as the origin of a circle.

Brightness-star association should be saved into a table (one or multiple) for all or at least a few stars. Then for a selected star, this brightness should be plotted as a function of time, or it could be also plotted into a histogram.

Input data

Input data are provided at the following address:

The directory is password protected. Please contact Michal Vrabel for access credentials.

Some steps of the work

  1. Review of methods for astrometric solving. How is it done in the world?
  2. Experiment with the astrometric solving method, including existing methods. Attempt to associate stars manually, try simply overlaying catalogue over the image and visualizing results. What problems are you seeing? How much is the lens of the AMON ACC camera distorting the image?
  3. Experiment with mitigating the deformations of the image if necessary.
  4. Implement code that realizes the astrometric solving for the whole dataset and store star brightness values in tables. If possible, store the absolute rotation of the fitted sky map. This might be useful in the process of creating the airglow image. Visualize the results and store the associations in tables.
  5. Experiment with creating the airglow images – consider the ideal method of determining the rotation angle of an image, consider the impact of simplification by rotating only using a simple transform in 2D, not the rotation of a sphere in 3D.
  6. Implement code that generates the airglow images for the whole dataset.
  7. Visualize the pixel brightness as a function of time for selected stars. Visualize the pixel brightness in a histogram for selected stars.

Computational resources

A remote computer for long-running code can be provided to students after discussion.

Semester 2020/2021

This project is assigned to Ján Korkoš and Vladimír Jacko.


Some star-matching related articles:

Maybe useful: