A Simplified Phase Display System for 3D Surface Measurement and Abnormal Surface Pattern Detection

[摘要] Today’s engineering products demand increasingly strict tolerances. The shape of a machined surface plays a critical role to the desired functionality of a product. Even a small error can be the difference between a successful product launch and a major delay. It is important to develop tools that confirm the quality and accuracy of manufactured products. The key to assessing the quality is robust measurement and inspection tools combined with advanced analysis. This research is motivated by the goals of 1) developing an advanced optical metrology system that provides accurate 3D profiles of target objects with curvature and irregular texture and 2) developing algorithms that can recognize and extract meaningful surface features with the consideration of machining process information.A new low cost measurement system with a simple coherent interferometric fringe projection system is developed. Comparing with existing optical measurement systems, the developed system generates fringe patterns on object surface through a pair of fiber optics that have a relatively simple and flexible configuration. Three-dimensional measurements of a variety of surfaces with curvatures demonstrate the applicability and flexibility of the developed system. An improved phase unwrapping algorithm based on a flood fill method is developed to enhance the performance of image processing. The developed algorithm performs phase unwrapping under the guidance of a hybrid quality map that is generated by considering the quality of both acquired original intensity images and the calculated wrapped phase map.Advances in metrology systems enable engineers to obtain a large amount of surface information. A systematic framework for surface shape characterization and abnormal pattern detection is proposed to take the advantage of the availability of high definition surface measurements through advanced metrology systems. The proposed framework evaluates a measured surface in two stages. The first step focuses on the extraction of general shape (e.g., surface form) from measurement for surface functionality evaluation and process monitoring. The second step focuses on the extraction of application specific surface details with the consideration of process information (e.g., surface waviness). Applications of automatic abnormal surface pattern detection have been demonstrated.In summary, this research focuses on two core areas: 1) developing metrology system that is capable of measuring engineered surfaces accurately; 2) proposing a methodology that can extract meaningful information from high definition measurements with consideration of process information and product functionality.

[发布机构] University of Michigan

[效力级别] high definition measurement

[关键词] surface measurement;high definition measurement;abnormal pattern detection;Mechanical Engineering;Engineering;Mechanical Engineering