Astrometry is the study of the position, distance, and motion of celestial objects. Global astrometry aims to connect stars in a large network to minimize position errors, while differential astrometry focuses on the relative positions of objects. The need for improved relative astrometric precision has been recognized by ESA's committee for Voyage 2050, leading to the proposal of the Theia mission with a 1-metre telescope capable of achieving sub-microsecond angular accuracy. Such extreme precision makes it possible to glimpse the architectures of the nearest exoplanetary systems, down to the mass of the Earth. The research discussed in this presentation focuses on the challenges of high-precision space astrometry missions and the calibration of the focal plane. Previous studies have shown that 1e-5 pixel precision is achievable with small detector arrays, but this research aims to validate this approach using larger detectors that would simplify the focal plane architecture. We plan to demonstrate the viability of large-format detectors (50 to 200 mega-pixels) in a laboratory test bed to ensure that they meet the required specifications. To solve the telescope stability issue, stars in the astrometric field can be used as metrological sources to calibrate the field distortion function at the required accuracy of 1e-5 pixels. The challenge now is to confirm this level of performance experimentally in the laboratory.