There is currently a strong push towards infrared astronomy, like the ground-breaking JWST and the upcoming ROMAN and Gaia NIR missions. The japanese JASMINE telescope will be the first near infrared (NIR) astrophotometric mission to focus on the Galactic central region and, in many senses, it will be pioneering the field of NIR high-precision astrometry for Milky Way (MW) dynamics. In order to test our data processing pipelines, we require a robust and reliable way to generate mock images. In this contribution, we present the JASMINE input catalogue: the most complete census of point-like sources in the NIR towards the Galactic centre. We used this catalogue as a blueprint from which to generate mock sources that resemble real stars as much as possible, while offering also the possibility of generating entirely new sources to compensate for the observational incompleteness. The method, while conceptually simple, requires treating each star of the input catalogue as new evidence that updates our prior knowledge, which in this case is represented by the underlying model of the MW used. The result is a custom probability distribution function for each star from which to draw mock sources. This represents the biggest and most realistic mock catalogue of the MW centre to date. In the future, we will improve it by adding more proper motions and parallaxes to the input catalogue, and by modelling the dependence of the distance on the kinematics.