Python analysis of the VVV Survey: a tool to prepare the EUCLID space mission

Jean-Baptiste Marquette

Audience level:


VVV is an ESO VISTA 4m telescope survey of the Galactic Bulge and Disk in YZJHK. Multi-epochs are being obtained in K band over several years. ESO is distributing the different data products, images and catalogues. These data are key to be able to select the optimal EUCLID microlensing fields towards the Galactic Bulge. We will present our new automated procedures to analyse these public data.


We developed a procedure to routinely extract stamps of all publicly available VVV images on selected fields. As an example, we choose fields with known planetary microlensing events. We stack them to get the highest signal to noise ratio image, and also built the K light curves of the chosen planetary microlensing events. It provides K band light curves, and when combined with microlensing modelling, estimates of the JHK baseline magnitudes. From a given target name and RA/DEC, different steps are performed in sequence: - query of the VVV survey through the ESO Data Portal to download all public data on a local storage RAID cluster (done once after each new data release) - creating a table of RA/DEC min/max (done once to update data after the above step) init of a configuration file (one per target) with some user-defined parameters - query of the 2MASS survey within a given box around the target to get reference stars fainter than 12 mag - identification of all relevant images from the target coordinates - extraction of stamps from the images with a given angular size centered on the target (when possible) - systematic PSF spatial variations calculations across the stamps - systematic PSF photometry using the above PSF models - automatic determination of the zero points, using Gaussian Mixture Models (GMM) and the convergence of the Bayesian Information Criterion (BIC), by matching the filed objects with the corresponding 2MASS reference stars, classified in the most relevant clusters in the diagram instrumental offset vs. 2MASS magnitude - extraction of the light curve of the target from the above photometric tables - stacking of the stamps to get the best S/N co-added image