Pattern recognition of rough surfaces by using goniometric scattered light

Abstract

Nowadays, most techniques for evaluating rough metal surfaces are based on tactile or confocal measurement procedures. However, these technologies have disadvantages in respect to measuring speeds, resistance to vibration, impact and dust. In this paper we present a novel surface measurement approach, which uses the scattering light technology. Our approach enhances the state-of-the-art scattering light-based surface measurement methodology in both the detector setup and evaluation of the raw intensity values acquired by the scattered light device. The main goal in optimizing the measurement setup is to capture scattering parameters for rough surfaces in a range greater than 10 µm based on an enlarged detector array. Regarding the evaluation, we propose a pattern recognition approach which maps the reflection intensity I back to material structures and the ten-point mean roughness Rz, the golden standard in tactile roughness characterization. Based on this approach, we are able to classify rough surface deviations like stripes using a simple but robust thresholding. In order to demonstrate the generality of our approach, we evaluate our approach using two rather different materials, i.e. brushed stainless steel and anodized aluminium.