Astronomical image processing from large all-sky photometric surveys for the detection and measurement of type Ia supernovae

Abstract

"Detection of transient events has become an important research subject in today's astronomy. To detect, report and study such phenomena, different informatics approaches have been proposed, among the most important of these are the image processing pipelines. Using the LSST Science Pipelines Stack (or Stack for short), a framework created by the Data Management Team of the Large Synoptic Survey Telescope, we have developed additions to one of these pipelines, focused on supernovae detection on the images from the Canada France Hawaii Telescope. We were able to run a complete pipeline using as input pre-calibrated exposures, performing an image subtraction and then select high quality candidates to be supernovae and transients. We obtained reasonable processing times by parallelizing most stages in the pipeline, and validated the Supernovae-Ia detection using data from the Supernovae Legacy Survey. Finally, we show a reduction of the overall number of source detections up to 80\% the amount in the base pipeline, and we report up to 95\% less light curve candidates, while preserving up to 85\% of Supernovae Ia with high signal present on the same period of time. We also present a simple method to label each detection per object, that allow us to show that the final light curve candidates have a high proportion of positive residuals which can greatly help other transient classification methods."--Tomado del Formato de Documento de Grado.

"La detección de objetos transitorios se ha convertido en un tema importante de investigación en la astronomía de hoy. Para detectar, reportar y estudiar dichos fenómenos, diferentes aproximaciones han sido propuestos, entre ellos, el uso de pipelines de procesamiento de imágenes. Usando el LSST Science Pipelines Stack (o Stack), un framework creado por el equipo de datos del LSST, desarrollamos adiciones a uno de dichos pipelines encargado de la detección de supernovas de tipo IA en las imágenes del telescopio CFHT. Logramos correr un pipeline completo, con tiempos reducidos de procesamiento, logrando una reducción de hasta 80% del número de detección dudosas, reportando 95% curvas de luz candidatas y preservando hasta el 85% de las supernovae Ia con alta señal presente."--Tomado del Formato de Documento de Grado.