Presentation Information
[1C03]The automation of radiographic image evaluation by applying artificial intelligence
○Yuichiro Kuroboshi1 (1. Osaka International Refining Company, Ltd.)
Keywords:
Artificial Intelligence,Radiography Test,Preventive maintenance
In refineries, Radiography Test (RT) is often used on pipes to evaluate their life. A wide range of cross-sectional wall thickness is measured quantitatively by using RT. Thickness reduction that is not in the cross section appears as shades on the image. Therefore, shades are important information for understanding the thickness reduction of piping. However, experience is required to detect shading visually and some troubles have occurred in recent years due to misjudgment of shading by inspectors. We introduce a case where AI successfully detected shades automatically.