Poster Abstract

P.48 Nathan Miles (Space Telescope Science Institute)

HSTCosmicrays: A python-based package for analyzing cosmic rays

HSTCosmicrays is a python-based pipeline designed to find and characterize cosmic rays found in dark frames (exposures taken with the shutter closed). Dark exposures are obtained routinely by all the Hubble Space Telescope (HST) instruments for calibration. The main processing pipeline runs locally or in the cloud on AWS by utilizing the astroquery package to query the Mikulski Archive for Space Telescopes (MAST) or the AWS HST Public Dataset, respectively. To date, we have characterized more than 1.3 billion cosmic rays in ~75,000 dark frames obtained with CCDs from the four active instruments STIS, WFC3/UVIS, ACS/HRC, ACS/WFC and the retired WF/PC2.

For the active instruments, the calibration pipelines identify pixels affected by transient events (including cosmic rays) and store this information in the data quality (DQ) arrays associated with each image. Legacy instruments do not have DQ extensions, so we employ a binary thresholding algorithm to identify all pixels that are statistically significant outliers (cosmic rays and hot pixels). For all instruments we perform a connected-component labeling analysis to identify groups of pixels affected by the same cosmic ray event and create a segmentation map for each image. We apply the segmentation map to the science extensions to extract a variety of morphological parameters for each individual cosmic ray. We use the dask parallelization framework to optimize each of the processing step. Finally, we use the HDF5 file format to store the extracted parameters for each cosmic ray found in an image, as well as a variety of image metadata.