Computer vision for automatic accurate plasma arc cutting measurement

Plasma arc cutting is a complex process with substantial interaction between process parameters. These manifest in deviations in the properties of the cut such as kerf width and kerf angle. Developing relationships between process parameters and aspects of the cut requires a large amount of data. This paper presents a novel method that applies computer vision to assess the kerf width against plasma cutting NC Code. The proposed method was applied to 25 cut parts and measured kerf widths to an accuracy of 0.2mm and was within 0.1mm of measurements made with a caliper. Measurements with this method were substantially quicker and provided more data than measurements made with calipers. No relationship was observed between kerf width and radius of curvature of the test cut.