Rapid detection of hairline cracks on the surface of piezoelectric ceramics

Abstract

Lead zirconate titanate, also known as PZT, is a type of piezoelectric ceramics commonly used for actuators in modern hard disk drives (HDDs). These PZT actuators are prone to hairline surface cracks, prompting detection and removal during the HDD production. Machine vision is then utilized for automatic detection of these cracks. The developed image processing approach comprises three steps: extraction of the region of interest, enhancement of crack regions, and elimination of irrelevant features. The key step, crack region enhancement, employs image filtering with a specifically designed filter kernel, capable of extracting thin crack regions from the rough surface of PZT actuators. The experiments show that the algorithm reveals cracks with high accuracy and high sensitivity, whereas the overall processing time satisfies the industrial environment.