About this Abstract |
Meeting |
13th International Conference on the Technology of Plasticity (ICTP 2021)
|
Symposium
|
13th International Conference on the Technology of Plasticity (ICTP 2021)
|
Presentation Title |
Image Feature Analysis for Magnetic Particle Inspection of Forging Defects |
Author(s) |
Quang-Cherng Hsu, Rui-Hong Ni, Jhan-Hong Ye |
On-Site Speaker (Planned) |
Quang-Cherng Hsu |
Abstract Scope |
Magnetic particle inspection uses the magnetic leakage phenomenon caused by defect of material that is used to detect the surface and subsurface defect of ferromagnetic material. Traditional inspection method will result in unstable and objective inspection problems; therefore, the development of automatic optical inspection is necessary. In this study, the forging constant velocity joint (CVJ) is used as the inspection sample. An aluminum alloy semi-circular diffuse cover of high-power ultraviolet light was proposed to reduce the reflective properties of inspection sample. A high-resolution CMOS industrial camera was used to capture magnetic particle image. For further automatic optical inspection, several image processing technologies were applied such as color separation, morphological operation, and contour detection. Then, all information of each possible crack defect such as area, perimeter, compactness, average width and length ratio can be obtained. Artificial neural network models were built to classify to defects and non-defects of forging parts. |
Proceedings Inclusion? |
Definite: At-meeting proceedings |