A virtual plane perpendicular to the axis was proposed, and the diameter of the measured axis was obtained quickly by using the least square circle fitting method. The linear structured light vision system was calibrated to obtain the camera’〖KG-*2〗s internal parameters, distortion coefficient and plane parameters of the structured light. The direction vector of the axis of the measured rotation body and the intersection point of the axis with the structural light plane were obtained, and the obtained direction vector of the axis and the intersection point of the axis with the structural light plane were used to establish a virtual plane perpendicular to the axis. The intersection points of the structure light plane and the surface of the part were projected onto the virtual plane, and the noise points in the data were eliminated by the distance from all projection points to the axis and the intersection points of the light plane. Finally, using the effective projection point and the distance between the axis and the intersection point of the optical plane, the cylindricity of the outer surface of the measured rotating body was obtained. The proposed algorithm was verified through the axial measurement experiment. In the complex measurement environment, when the diameter of the shaft under test is 30~50 mm, the average measurement error is 30 μm and the maximum measurement error is 50 μm. Therefore, the proposed algorithm can be applied to the diameter measurement of practical rotary parts.