Machine vision automates inspection of thick-film hybrids

The function of a machine vision system for inspection is discussed, and its image acquisition and image processing aspects are examined in detail. It is concluded that the image-acquisition system must be optimized for the specific optical characteristics of the circuit substrate in order to obtain a high-contrast image. The image processing section must be able to correctly segment the image into the specific layers. The defect analysis must be able to identify flaws one-tenth the size of the smallest feature of the circuit, while not flagging acceptable anomalies as defects.