2 Si3N4 ceramic bearing inner
ring surface defects identification and classification platform
construction
The prerequisite for analyzing and predicting surface defects inside the
inner ring of Si3N4 ceramic bearings
using deep learning models is a large-scale dataset. To produce a
Si3N4 ceramic bearing inner ring surface
defect dataset, the surface defect images are collected by building a
non-destructive inspection platform. The inspection platform is composed
of feeding module, image acquisition module and image processing module.
The feeding module includes mechanical arm, coaxial rotating gear and
motor. The image acquisition module consists of high-speed CCD camera,
stereo microscope and LED light source. The image processing module
includes image acquisition card, image memory card and industrial PC.