Pointing correction#

Star tracker method#

The star tracker method estimates telescope pointing corrections using variance images computed from interleaved sky pedestal events. Stars in the field of view produce localized increases in the pixel variance. By reconstructing the positions of these stars and matching their observed trajectories to a stellar catalog, the pointing offset of the telescope can be inferred.

The pointing correction calculation for the star tracker method is performed using the calibpipe-star-tracking tool.

The following configuration parameters are required:

  • TableLoader.input_url: Input containing the variance images derived from sky pedestal events.

  • HDF5MonitoringSource.input_files: List of input files for the monitoring source.

  • output_path: Output filename.

StarTrackerTool orchestrates the reconstruction of stellar trajectories in the Cherenkov camera and the estimation of pointing corrections. It requires the following fields:

  • chunking_type: Name of the Chunking subclass used to group variance images for the reconstruction.

The star tracker method is composed of the following components:

  • StarFieldQuery: Retrieves stars from a catalog for the nominal telescope pointing direction.

  • StarFieldReconstructor: Reconstructs stellar positions in the camera from the variance images.

  • StarFitter: Uses orthogonal distance regression (ODR) to fit the expected star positions to the reconstructed stars and determines the pointing offset.

Additionally, the user must configure the selected chunking strategy and, optionally, the point spread function (PSF) model used for the star reconstruction.

StarFieldQuery#

The StarFieldQuery component retrieves stars expected to be visible in the field of view.

It requires the following fields:

  • catalog_name: Name of the stellar catalog used for the query (e.g., “Hipparcos” or “Yale”).

  • apply_proper_motion: Whether stellar proper motion corrections are applied.

  • region_radius: Angular radius around the nominal pointing direction used for the catalog query.

  • max_magnitude: Maximum stellar magnitude included in the reconstruction.

  • wavelength_band: Magnitude band used for star selection.

StarFieldReconstructor#

The StarFieldReconstructor component identifies star candidates in the variance images and reconstructs their positions.

It requires the following fields:

  • psf_model_type: Name of the PSF model used to describe stellar images in the camera.

  • pdf_bin_range_factor: Range factor used for the construction of the star detection probability density functions.

  • n_pdf_bins: Number of bins used in the probability density functions.

  • pdf_percentile_limit: Lower percentile threshold used to suppress background fluctuations.

  • min_star_prominence: Minimum prominence required for a star candidate to be considered in the reconstruction.

PSF Models#

The reconstruction of stellar images relies on a camera-specific point spread function model.

The following implementation is currently available:

  • ComaPSFModel: Analytical description of the asymmetric optical point spread function including radial and azimuthal distortions across the camera field of view.

The model parameters are telescope-dependent and are typically derived from dedicated optical calibration studies.

StarFitter#

The StarFitter component associates reconstructed stars with catalog stars and determines the pointing correction using ORD.

It supports the following configuration field:

  • exclude_invalid_stars: Ignore stars in the fit which reconsctruction failed for at least one of the variance images. It only includes stars that have a valid reconstruction over the whole set of variance images in the fit.

Chunking#

Variance images are processed in groups to improve the robustness of the trajectory reconstruction.

The following chunking implementation is available:

  • SizeChunking: Creates chunks containing a fixed number of variance images.

The SizeChunking component supports the following fields:

  • chunk_size: Number of variance images per chunk.

  • allow_undersized_tables: Allow processing of datasets smaller than the requested chunk size.

Below is an example configuration for estimating telescope pointing corrections from variance images of interleaved sky pedestal events.

# Configuration for the calculation of the pointing corrections via the StarTracker method.
StarTrackerTool:
  chunking_type: "SizeChunking"
StarFieldQuery:
  catalog_name: "Hipparcos"
  apply_proper_motion: True
  region_radius: 2.0 degree
  max_magnitude: 7.0 mag
  wavelength_band: "BTmag"
StarFieldReconstructor:
  psf_model_type: "ComaPSFModel"
  pdf_bin_range_factor: 0.1
  n_pdf_bins: 25
  pdf_percentile_limit: 0.001
  min_star_prominence: 3
# LST-1 configuration for the Coma PSF model
ComaPSFModel:
  asymmetry_max: 0.49244797
  asymmetry_decay_rate: 4.723892639969285
  asymmetry_linear_term: 0.07782730217679201
  radial_scale_offset: 0.02755256784329137
  radial_scale_linear: 0.030690191771900095
  radial_scale_quadratic: 0.030690191771900095
  radial_scale_cubic: -0.007768185644664285
  polar_scale_amplitude: 0.48543114
  polar_scale_decay: 3.862262965579431
  polar_scale_offset: 0.039844600455081855
StarFitter:
  exclude_invalid_stars: True
SizeChunking:
  allow_undersized_tables: True
  chunk_size: 25