An improved methodology of melt pool monitoring of direct energy deposition processes
Robert E. Sampson, R.J. Lancaster, Mark Sutcliffe, David Carswell, Carl Hauser, Josh Barras
Abstract
Additive manufacturing processes have previously benefited from the introduction of melt pool dimensioning systems. These typically measure melt pool width by performing binary thresholds and highlighting edges using common edge detection algorithms. Melt pool monitoring systems have been successfully used to develop control systems and enhance process understanding. This paper presents an improved machine vision technique to enhance images in melt pool monitoring systems. Enhanced images contain features that indicate true melt pool edges. The research highlights potential flaws in more established emissivity-based image processing algorithms and a new image processing technique is developed. The new technique produced improved accuracy and performed melt pool measurements independent of emissivity values.