The objective of this research is to advance computer vision techniques
and their applications in the electronics manufacturing industry. The research has
been carried out with specific reference to the design of automatic optical inspection
(AOI) systems and their role in the manufacture of printed circuit boards (PCBs).
To achieve this objective, application areas of AOI systems in PCB manufacture
have been examined. As a result, a requirement for enhanced performance
characteristics has been identified and novel approaches and image processing algorithms
have been evolved which can be used within next generation of AOI systems.
The approaches are based on gaining an understanding of ways in which
manufacturing information can be used to support AOI operations. Through providing
information support, an AOI system has access to product models and associated
information which can be used to enhance the execution of visual inspection
tasks. Manufacturing systems integration, or more accurately controlled access to
electronic information, is the key to the approaches. Also in the thesis methods are
proposed to achieve the flexible integration of AOI systems (and computer vision
systems in general) within their host PCB manufacturing environment. Furthermore,
potential applications of information supported AOI systems at various stages of
PCB manufacturing have been studied.
It is envisaged that more efficient and cost-effective applications of AOI
can be attained through adopting the flexible integration methods proposed, since
AOI-generated information can now be accessed and utilized by other processes.

On flexibly integrating machine vision inspection systems in PCB manufacture

Many defects in electronic products can escape detection by some of the most sophisticated electrical testing and vision inspection equipment available for manufacturers’ use.

Ronald Pound

Papers

Image Processing and Non-Destructive Testing