City Research Online

Abstract

This thesis is concerned with the application of modern concepts in instrumentation, and particularly, signal processing, to systems for intelligent industrial automation. Its chief objective is to provide a signal processing methodology for defect detection in automated visual inspection.

Previous achievement in the field is surveyed in the introduction. Since the properties of signals used in surface inspection depend critically on the opto-electronic scanning devices used for signal acquisition, these receive first consideration. Then, from concepts generated originally for target detection in radar and sonar, a canonic form is developed for the detection process. This comprises the three consecutive stages, contrast enhancement, decision, and trigger association. Alternative forms of processing for the three stages are proposed, and analysed theoretically using statistical concepts. Computational experiment, in which the processing methods are implemented as computer programs operating on stored data, is then used, to confirm the practical effectiveness of the methods, and to compare alternatives. A distinction is made between local and global defects, and methods are discussed for detecting both kinds. Alternative methods of combining the various methods together are described and compared. Finally, a complete system is proposed, for the detection of visually perceivable defects on cold rolled steel strip, which is the material used during the simulation experiments. It is concluded that the suggested processing will yield inspection systems whose performance meets or exceeds operational requirements, and is constantly superior to that of a human operative in speed of operation, consistency, objectivity, and in ability to detect defects of poor contrast.

Publication Type:	Thesis (Doctoral)
Subjects:	H Social Sciences > HD Industries. Land use. Labor > HD28 Management. Industrial Management Q Science > QA Mathematics > QA75 Electronic computers. Computer science T Technology > TJ Mechanical engineering and machinery
Departments:	School of Science & Technology > School of Science & Technology Doctoral Theses Doctoral Theses

Export

Downloads