Machine vision based inspection of oil seals

In this research, the dimensions (inner and outer diameters) and surface defects of oil seals are measured or inspected by a machine vision system. There are four steps of the measurement: (1) obtain the threshold value of an image, (2) use that value and a boundary-tracking method to ascertain the positions of boundary points, (3) use the Gaussian technique to smooth the boundary points, and (4) use the minimum errors area method to calculate inner and outer diameters. Before inspecting for surface defects, the center position must be calculated. With that position determined, the gray levels of the circumferential pixels in the inspected regions can be obtained. To check the gray levels, an improved moving-range chart and an individual observation control chart are employed, thus completing the inspection.