Third order loose coupled hybrid intelligent system for machine vision applications

Increased use of machine vision systems is making a significant contribution to ensuring competitiveness in modern manufacturing. However, the multi-disciplinary nature of machine vision necessitates a high level of competence in a diverse range of subjects. These factors have resulted in a shortage of qualified machine vision engineers. Overcoming this skills shortage can be approached in two ways; the education of more machine vision engineers or a reduction in the time taken to design and implement a machine vision system. One possible solution to both these problems is the use of an intelligent design tool that will aid an engineer in designing a complete machine vision system. This work focuses on one component of such a tool, a hybrid intelligent system to aid in the selection of algorithms for use in a particular machine vision task. The approach taken in this work belongs to the more general category of systems for the intelligent supervision of a library of operators and as such can be generalized to other applications and domains. The system design is based around an evolutionary algorithm (EA), case based reasoning (CBR) and rule based reasoning (RBR) architectures.