ABSTRACT
A generally accepted definition of machine vision is “…the analysis of images to extract data for controlling a process or activity.” Machine vision is a subdiscipline of artificial intelligence that uses video cameras or scanners to obtain information about a given environment, and to extract data from digital images about objects in the image. Machine vision takes an image in and outputs some level of description about the objects in it (i.e., size, position, correlation, etc.). Although applications for machine vision vary, there are generally only a few main categories:
• Quality assurance • Sorting • Material handling • Robot guidance • Calibration
Quality assurance represents probably the largest number of machine vision applications in use today. All that quality assurance machine vision systems do is inspect an image and decide whether the part should pass or fail, based on a predetermined set of rules. Although humans can be excellent quality assurance inspectors, vision systems are appropriate when images need to be inspected at high speed (often production lines require part inspection rates up to many complex parts per second), and for monotonous work. Vision systems do not become bored, or lack concentration; therefore, quality assurance can be much more highly controlled. Conversely, human inspectors can think and possibly expand or change the pass/fail rules-this can be both advantageous and disadvantageous. If the human finds a flaw in the product that is not part of the pass/fail criteria, then it can be added. On the contrary, humans can interpret rules incorrectly, often .leading to quality assurance issues.
