Qualification of AI-Based Surface Topography Inspection Systems for Inline Measurement in Series Production: Tactile Touch Systems Versus Optical AI Analysis

Surface evaluation is a critical aspect in determining the reliability and quality of technical products, particularly those with functional or safety-critical features. This study focuses on surface characteristics, including geometric shape, and roughness, emphasizing their significance in diverse industries such as automotive. Key parameters for workpiece profile analysis, such as form deviation, surface topography, and length are explored. Traditional stylus methods, involving a stylus tip traversing the surface, are conventionally employed for recording characteristic values related to form deviation or roughness profiles. This paper introduces artificial intelligence (AI) methods as an alternative for surface topography assessment, offering new possibilities in terms of defect detection, working speed, and cost-effectiveness compared to traditional methods. The AI-based measurement principle involves training an algorithm on a comprehensive set of images, enabling swift surface assessment in series production inspections. The focus is on developing a procedure for the qualification of AI-based surface inspection systems, exemplified by surface roughness for inline measurement in series production. A case study analyzing cutlery compares the effectiveness and efficiency of AI-based inspection to classic tactile methods, providing a comprehensive exploration of their respective advantages and disadvantages. Emphasis is placed on fine shape deviations, including roughness as well as the significance of roughness parameters and 3D measurement using confocal scanning.