iRVision GPM Locator Learning Feature

This iRVision tech short, presented by FANUC engineer Josh Person, focuses on utilizing the Geometric Pattern Match (GPM) locator tool in 2D and 3D applications. GPM uses an edge-based training algorithm to identify contrasting edges, requiring users to manually mask out unwanted features like glare, shadows, or background noise to ensure quality. Crucially, the video demonstrates the powerful GPM learning feature, which automates this masking process. By feeding the system multiple examples of the part in different orientations and with typical variations (like surface flaws), the learning algorithm automatically creates a robust model. This eliminates noisy edges, drastically improving the GPM score and adding robustness to the vision process, which is especially vital when dealing with less-than-perfect or non-precision manufactured parts.