This thesis addresses the need for adaptability in vision systems that measure system state information in a sensory feedback role for the control and coordination of flexible discreteitem materials handling operations, such as those performed by a robotic palletising system. In addition, this thesis addresses the need for vision systems that are more easily configured by users, such as factory technicians and operators, who have lower skill levels than those generally required to (re-)configure a machine vision system. In response, a unique coupling mechanism and intuitive human-computer interface have been developed, hiding the complexity of image analysis from the end-user and simplifying the way that a machine vision system is configured. The mechanism couples machine vision-related "visual checks” to materials handling tasks in a generic framework of materials handling activities. Visual checks which define what control information is required are implicitly linked to image analysis functions which define how that information is extracted from digitised images of a materials handling system. Consequently, this research has developed a set of task - visual check "building blocks" that can be used in various combinations to define the sequence of actions and image analysis required to perform a variety of materials handling operations. In addition, a number of pre-defined task – visual check combinations and mechanisms for manipulating them have been developed, providing solution templates that can be used immediately or modified to suit application-specific requirements. These developments have been realised together with several aesthetic, ergonomic and functional features in a machine vision configuration interface, known as SlyVision. SlyVision's modularity, extensibility and upgradeability expressed to both the end-user and the system developer through its underlying object oriented architecture and intuitive user interface design make important contributions to its overall adaptability. Demonstrations involving a typical palletising and a de-palletising operation have shown how SlyVlsion is used to specify visual checks and configure the associated machine vision components without requiring the end-user to select or apply image analysis techniques or functions. In addition, the relative simplicity of the configuration process is demonstrated. Consequently, these developments assist people with limited understanding of machine vision technology to set up and maintain a vision system, thereby improving their ability to keep pace with frequent changes in their materials handling operations, while limiting the cost in time, money and effort required to (re-)configure a vision system. / Whole document restricted, but available by request, use the feedback form to request access.
| Identifer | oai:union.ndltd.org:ADTP/247524 |
| Date | January 1997 |
| Creators | Sly, Ian M. P. |
| Publisher | ResearchSpace@Auckland |
| Source Sets | Australiasian Digital Theses Program |
| Language | English |
| Detected Language | English |
| Rights | Whole document restricted but available by request. Items in ResearchSpace are protected by copyright, with all rights reserved, unless otherwise indicated., http://researchspace.auckland.ac.nz/docs/uoa-docs/rights.htm, Copyright: The author |
Page generated in 0.002 seconds