A PROCEDURAL METHOD FOR GENERATING PBR TEXTURE MAPS FROM HISTORICAL GEOMETRIC PATTERNS

Abstract

This paper proposes a procedural method for generating PBR texture maps from a parametric geometric model of a historical pattern. A single geometric mask is used as a common spatial source for the Base Color, Height, Normal, and Roughness maps, while surface micro-variations are represented by a separately controlled deterministic function. A reproducible 1024 × 1024 prototype was implemented using Python and NumPy to verify cross-map geometric consistency, normal-vector normalisation, and sensitivity to changes in a geometric parameter. The experimental results confirm the consistency and functionality of the proposed computational pipeline, while further physical calibration is required before the generated maps can be interpreted as representations of measured properties of historical materials.