The texturing of steel surfaces is well known to affect their functional performance. Different processes (e.g. forming, painting) are affected by different wavelengths of surface structures which, in order to study the phenomena, are usually separated through filtering processes. New texturing processes call for new filtering techniques, and amongst the latest ones, the Gram-Schmidt (GS) decomposition technique proved to be suitable for deterministic patterns such as Electron-Beam Textured (EBT) surfaces.Optimisation lead to new textures, with more dense patterns, so extending the GS methodology seemed natural and meaningful. This paper describes how the technique was algorithmically adapted and how it will benefit process control.