Design and Development of Non-Contact Image Processing Technique to Monitor Surface Texture During Turning

Abstract

Surface roughness is a critical quality indicator in machined components, directly influencing fatigue life, wear resistance, and functional performance. This study presents an image-based MATLAB code for estimating roughness parameters—Ra (Arithmetic Average Roughness), Rq (Root Mean Square Roughness), and Rz (Ten-point Mean Roughness)—from microstructure images under various tool rotational speeds and traverse speeds. The outputs are validated against experimental profilometry data across nine parameter sets. While trends in Ra, Rq, and Rz are generally well captured, discrepancies increase with surface complexity. The maximum deviation observed in Ra was 2.74 µm, corresponding to a high-feed, high-speed condition. Despite overestimations in most cases, the MATLAB method provides a fast, non-contact estimation approach for comparative roughness evaluation. This image-processing-based approach holds promise for rapid surface quality assessments in manufacturing environments.