Parallel Light Source Based Machine Vision for Length Measurement of Shaft Part Chamfer

Chamfering processing on shaft parts is an essential step for production and assembly. This paper studies the online chamfer length measurement of shaft parts by machine vision, aiming to establish an automatic machining quality ensuring system. The challenge of the measurement is that there is various interference on the cylindrical surface of shaft parts before final processing. If conventional imaging method is utilized, the interference may show visual characteristics similar to chamfer, leading to misjudgment. To solve this issue, we proposed a parallel light source based imaging method, which can illuminate the chamfer surface and suppress the reflected light from other areas. Additionally, a weighted least squares fitting based chamfer width calculation is presented to avoid the influence of iron filings and burrs. In experiments, the advantages of our technology in interference resisting are illustrated through comparison. The abnormal chamfer dimension on several shaft samples is successfully detected.