No abstract

No abstract

No abstract

While CNC programmers have powerful tools to develop optimized toolpaths and machining plans, these efforts can be wholly undermined by something as simple as human operator error during fixturing. This project addresses that potential operator error with a computer vision approach to provide coarse, closed-loop control between fixturing and machining processes. Prior to starting the machining cycle, a sensor suite detects the geometry that is currently fixtured using computer vision algorithms and compare this geometry to a CAD reference. If the detected and reference geometries are not similar, the machining cycle will not start, and an alarm will be raised. The outcome of this project is the proof of concept of a low-cost, machine/controller agnostic solution that is applied to CNC milling machines. The Workpiece Verification System (WVS) prototype implemented in this work cost a total of $100 to build, and all of the processing is performed on the self-contained platform. This solution has additional applications beyond milling that the authors are exploring.

No abstract