Abstract
Surface topography is a critical quality characteristic of many products and manufacturing processes. Various defects commonly appear on the topography of finished products in spatial patterns after or during manufacturing. These defects are difficult to be identified using traditional monitoring approaches because of the complex structure of topographic data. This article develops a novel and effective approach for monitoring spatial defects in topographic surfaces. The approach improves the representation of surface characteristics through the developed multilabel separation-deviation surface (MSS) model, which labels the important surface characteristics and smooths out the noisy characteristics. We develop two features for monitoring changes in the characteristics of the assigned labels. The MSS feature is introduced for capturing deviations within the assigned labels, and the generalized spatial randomness feature is derived for quantifying deviations between the assigned labels. These two features are integrated into a single monitoring statistic, which is successfully applied for detecting various defects in topographic surfaces, outperforming the traditional monitoring approaches.