AME - Event camera implementation for manufacturing in-situ diagnostics

Event cameras (https://en.wikipedia.org/wiki/Event_camera) use an imaging sensor that responds to local changes in brightness rather than recording the irradiance.  Rather than using a shutter, each pixel asynchronously reports changes.  This is based on the way our retinas respond to light and is sometimes called a neuromorphic camera.  This project aims to implement an event camera in laser powder bed fusion to log changes around the melt pool and capture splatter.  These attributes can be used to generate a voxelized map that highlights potential defects in parts.  This has been done by our research group and others using conventional cameras.  The advantage is that the event camera can achieve a high dynamic range with an effective frame rate greater than 50,000 frames per second.  The goal is to build on initial work at Los Alamos on using event cameras for process monitoring.

We will purchase a moderate-cost event camera.  Students will learn to work, processing data streams into actionable images.  This will involve adapting open-source code.  Of particular interest is integrating the camera with laser speckle techniques for understanding geometric changes within a scene.  After developing a working knowledge of the event camera, we will record movies of the laser powder bed fusion process and correlate the data with the process conditions. Other applications can also be considered based on student interest.

The project will involve an initial literature review to select the camera, basic optical setup, and coding to interpret the results.