The extreme extinction (A_V_~30mag) and its variation on arc-second scales towards the Galactic centre hamper the study of its stars. Their analysis is restricted to the near infrared (NIR) regime, where the extinction curve can be approximated by a broken power law for the JHKs bands. Therefore, correcting for extinction at these wavelengths is fundamental to analyse the structure and stellar population of the central regions of our Galaxy. We aim at: (1) Discussing different strategies to de-redden the photometry and checking the usefulness of extinction maps to deal with variable stars. (2) Building and making publicly available extinction maps for the NIR bands JHKs. (3) Creating a de-reddened catalogue of the GALACTICNUCLEUS (GNS) survey, identifying foreground stars. (4) Performing a preliminary analysis of the de-reddened $K_s$ luminosity functions (KLFs). We use photometry from the GNS survey to create extinction maps for the whole catalogue. We take red clump (RC) and red giant stars of similar brightness as a reference to build the maps, and de-redden the GNS photometry. We discuss the limitations of the process and analyse non-linear effects of the de-reddening. We create high resolution (~3'') extinction maps with low statistical and systematics uncertainties (<~5%), and compute average extinctions for each of the regions covered by the GNS. We check that our maps effectively correct the differential extinction reducing the spread of the RC features by a factor of ~2. We assess the validity of the broken power law approach computing two equivalent extinction maps A_H_ using either JH and HKs photometry for the same reference stars, and obtain compatible average extinctions within the uncertainties. Finally, we analyse de-reddened KLFs for different line-of-sights and find that the regions belonging to the NSD contain a homogeneous stellar population that is significantly different from the one in the innermost bulge regions.