Astreaks

Astreaks is a robust, modular, automated pipeline for Astrometry of Asteroids in Streaked Images. Astreaks is a package implemented for automated astrometry of non-sidereally tracked minor planets, where the reference stars have significant elongation.

As the proper motion of targets increases from a couple of arc-seconds to tens of arc-seconds per minute, non-sidereal observations lead to streaking background reference stars which makes the end goal of accurate astrometry very challenging. The existing source extraction software are not designed to detect sources with tens or hundreds of arc-seconds of elongation. Robustness of source extraction process to generate a catalogue of objects is of utmost importance in astrometry. The GROWTH-India astrometry pipeline for NEOs – Astreaks – achieves this goal by analysing calibrated non-sidereal data in following steps: sky background estimation, streak spread function (SSF) model generation for reference stars, thresholding, image segmentation and source deblending, synthetic image generation and getting astrometric solution, and getting astrometry and photometry of the targets.

Implementation

Astreaks is a python-based pipeline that relies on the standard modules in astropy, and other standard libraries like numpy, scipy and matplotlib. Here we describe the key steps in the processing pipeline:

Astrometry

This synthetic image can now be solved using any standard software. We use the astrometry.net offline engine to get the WCS solution of this image. This WCS is then used to get the positions of the minor planet at the mid-time of the exposure. The synthetic image is calibrated for photometry using the fluxes in the source catalog.

Hyperparameters, Parameters & Input Files

Inputs for Astreaks are divided into three broad categories:

• Hyperparameters: Values in the configuration file which have to be updated only when there is some change in the instrument being used – for instance pixel scale, meshgrid size, etc.

• Parameters: Values like target coordinates which have to be supplied for each file being solved

• Input files: The actual data – one sidereal image (for PSF extraction) and the target non-sidereal image

The astrometry configuration file for our pipeline comprise of several hyperparameters required at different stages of source extraction. Background estimation requires an appropriate mesh-size. Thresholding level and contrast level is specified for source deblending. SSF model generation requires the pixel scale. Synthetic image generation requires the size of the PSF model and to solve the synthetic image for WCS, the search radius needs to be specified depending on the field of view. For photometry, we need to specify the CCD gain, photometry distance threshold and circular and annular aperture size.

The parameters required for processing include tracking rate, velocity position angle, telescope position angle, approximate right ascension and declination for the target, and exposure time. These parameters are parsed from the image header, resulting in full-automation. These can be overwritten by explicitly parsing them in command line. The two input files needed include a sidereal image from the same night of observations (preferably, in the same field at nearly the same time) and a target_info file that maps the images to (x, y) coordinates of the NEO in the image, for which the MPC report file has to be produced.

Validation

First, astreaks was tested on non-sidereal data for 18 NEOs acquired with the Apogee camera. The Apogee KAF3200EB camera gives a 11' * 7.5' field of view and 0.307"/pix resolution when mounted on the GROWTH-India telescope. The NEOs observed had a proper motion in the range 3.85 - 120"/min. Exposures of 40 - 300s were taken, resulting in streaks of length 14.16 - 140.86". We obtained astrometry solution for 133 images of these 18 NEOs using our pipeline. Orbit fit was attempted by appending all existing observations at MPC, as of the date of our observation. The residuals obtained had a standard deviation of 0.58", which is respectable for astrometry using such extended sources.

Next we tested our pipeline on non-sidereal data for 32 NEOs obtained using another instrument set-up: Andor camera. The Andor iKon-XL 230 4k back-illuminated CCD is the primary camera of GROWTH-India Telescope and has a 0.7 degree field of view at 0.67"/pix resolution. By virtue of the modular nature of this package, the only tweak required to shift data processing from one instrument to the other was to update the hyperparameters in astrometry configuration file. The tracked NEOs had a proper motion in the range 0.99 - 29.9"/min. Exposures of 120 - 500s were taken, resulting in streaking reference stars of lengths in the range 8.25 - 228.58". Orbit fit was attempted by appending all existing observations at MPC, as of the date of our observation. The residuals obtained had an admirable standard deviation of 0.47".