DESIGN OF A CLOSE-UP STEREO VISION BASED SHEET METAL INSPECTION SYSTEM

This paper describes the design and implementation of a close-up stereo computer vision based sheet metal strain and geometry inspection system. A square grid pattern is first applied to the part using either electrochemical etching or screen printing. The part is formed, and measured using a two camera sensor that can be attached to a CMM or FARO arm. Image processing is used to extract the corners of the deformed grid for either three dimensional reconstruction, or for strain analysis. Graphical results are conveniently displayed using HOOPS. Sample measurements using both steel and aluminum test domes indicate an achievable accuracy comparable to other systems.