Vertical edge detection-based automatic optical inspection for solder jet ball joint defect on Head Gimbal Assembly

This paper describes an image processing method to automate the optical inspection process for solder jet ball joint defects on the Head Gimbal Assembly (HGA) of hard disk drives (HDDs). Vertical edge detection methods are proposed to identify the defect. The performance of the vertical edge detection method is evaluated against another technique; the Sobel-based method. Both methods were implemented and tested with several solder joints. The experimental results show the superiority of the vertical edge detection method, which had an accuracy of 99.32%, as compared to the Sobel based method, with an accuracy of 80.02%.