Stereoscopic Inspection for Microelectronics Using Multiple Imaging Sensors

Stereoscopic imaging is a powerful technique that can provide (1) 3D visual inspections (for both surface and interior features), (2) 3D measurements, and (3) 3D defect algorithms. We propose a PC-based microelectronics inspection system that takes advantage of not only stereo imaging, but also multisensor imaging. Method will be developed for obtaining real-time stereo images from X-ray and infrared cameras, both of which are very important for microelectronics. Real-time algorithms will be developed for automatic depth measurement using stereo pairs from a single sensor type. Then the depth information from two or more sensors will be used in conjunction to develop defect recognition algorithms that make use of the added information. We will explore how new signal processing and detection technologies (e.g., neural networks, fuzzy logic, and rule-based systems) can aid in automating microelectronic inspections. Our primary goals for Phase I are to demonstrate stereoscopic 3D imaging of typical microelectronic components on a PC computer, using both X-ray and infrared images, and to produce a System Concept Document that will (a) review current literature and inspection procedures, (b) explore concepts and algorithms for multisensor data fusion, and (c) present a preliminary design for the prototype system to be built during Phase II.