Vision inspection with k-means clustering on head gimbal assembly defect

Abstract

Head Gimbal Assembly (HGA) is an important feature of read and write process in a Hard Disk Drive (HDD). Currently, HGA circuit inspections are done using human operators under microscope; the vision processing for inspection in automated systems is required. This research work proposes an algorithm for detection the HGA circuit defect by using the blob detection, and then analysis the properties of blob tool. By the measurement properties of the blob tool, the K-Means Clustering can specify the data in each group in 95.45% accuracy with 110 samples.