T-PHOT: A code for (multiwavelength prior-based) template-fitting photometry


The advent of deep multiwavelength extragalactic surveys has led to the necessity for advanced and fast methods for photometric analysis. In fact, codes which allow analyses of the same regions of the sky observed at different wavelengths and resolutions are becoming essential to thoroughly exploit current and future data. In this context, a key issue is the confusion (i.e. blending) of sources in low-resolution images.
T-PHOT is a publicly available software package developed by Emiliano Merlin within the ASTRODEEP project. T-PHOT is designed to measure robust photometry from low-resolution images, where the blending of sources can be a though problem for the accurate and unbiased measurement of fluxes and colours.
In short, T-PHOT gathers data from a high-resolution image of a region of the sky, and uses this information (source positions and morphologies) to obtain priors for the photometric analysis of the lower resolution image of the same field. T-PHOT can handle different types of datasets as input priors, namely i) a list of objects that will be used to obtain cutouts from the real high-resolution image; ii) a set of analytical models (as .fits stamps); iii) a list of unresolved, point-like sources, useful for example for far-infrared (FIR) wavelength domains.
T-PHOT proves to be a state-of-the-art tool for multiwavelength optical to far-infrared image photometry. Given its versatility and robustness, T-PHOT can be considered the preferred choice for combined photometric analysis of current and forthcoming extragalactic imaging surveys. It will be included in the photometric pipeline for the analysis of the EUCLID ESA mission data.
Screenshot from 2018-02-21 22-26-24
T-PHOT workflow diagram
In the  public release v2.0  new features and options have been implemented:
• Background estimation, with two methods: global subtraction of a constant fitted value, and local fit of individual “background templates”
• Local/individual kernel fitting: it is possible to associate a different kernel to each source to optimize the fit coping with local variations
• Individual source registration (dance): after the fit, a refinement of the spatial registration of the objects is performed on individual basis rather than on arbitrary regions.
• Flux prioring: the measured flux of selected sources can be constrained to be within a given interval around a given input value
• Statistics on the residuals: the output includes a text file with diagnostic statistics for each source, based on the residual image produced after the fit
• RMS threshold to exclude sources from the fit: if a source includes a pixel with RMS exceeding a given value, it will be excluded from the fit.
 Production of model and residual images with selected sources: this option can be useful to isolate or subtract a selection of chosen objects.
The detailed description of the code, along with an extended set of tests, can be found in Merlin et al., 2015, A&A, 582, 15 and in Merlin et al., 2016, A&A, 595A, 97M.
Abell2744 cluster IR bands processed with T-PHOT using real priors for faint galaxies and models for cluster objects. Top to bottom: Hawk-I Ks, IRAC1 3.6, and IRAC2 5.0 bands. Left: original images; right: residuals after T-PHOT fitting. See Merlin et al. 2016, A&A 590A, 30M
[The Tea-Pot logo idea is ©David Elbaz. The logo was actually found on Google so if any copyright exists please contact me!]