IIC Journal of Innovation 19th Edition The Role of Artificial Intelligence in Industry | Page 18

Applicative Framework for End-to-End AOI Implementation
This approach prevents the assembly of a wrong component in the final product and thus reduces waste .

3.2 AOI FOR POROSITY DETECTION

The designed system has been implemented in a gear pump assembly line to detect porosity on the outer surface of the housing . The housing is manufactured using die-casting technology and subsequently performs a surface heat treatment that makes it black in color .
Porosity is the most common casting defect . If not detected during the assembly process , when the pump operates at full speed porosity can lead to fluid leakage and reduced performance .
Before implementation , line operators used to perform this check visually . However , visual check is commonly considered as not reliable 4 .
That is why the company decided to integrate the designed system in the existing line , exploiting its non-invasive nature : the final station , the one designated for the visual check , already featured a camera for DMC reading . It has been equipped with a new camera for image recording and a presence sensor , in order to replicate the design in Figure 2-1 .
Figure 3-4 : Top ( left image ) and bottom view ( right image ) of the designed solution applied to porosity detection use case . On the left , the gear pump is positioned by the operator on the control position , where the sensor detects its presence . On the right , the camera and the illumination system .
4
M . -J . -J . Wang and C . -L . Huang , ‘‘ Evaluating the eye fatigue problem in wafer inspection ,’’ IEEE Trans . Semicond . Manuf .
IIC Journal of Innovation 13