Machine vision

From ArticleWorld

Machine vision is an advanced field of image processing in which the concept of computer vision is applied in industry and manufacturing. With the interface of computer networks along with image sensing devices, robotic arms and other equipment can be controlled. Making the use of optics, computer engineering and automation, ‘seeing- robots’ are equipped.

Applications of a machine vision

The most common application of machine vision is in the inspection of manufactured goods for defects, incompletion etc. Microchips, automobiles, canned food and pharmaceutical items can be scanned using machine vision. Machine vision systems use smart cameras and image processing techniques to point out any discrepancies. Thus quality control checks which are normally carried out by humans can be done by such systems, with little chances of error. There is no question of distraction as is possible in human operators.

Counting objects, reading batch numbers and looking for external defects can thus be done on conveyer-belt or assembly line arrangements with high repeatability of operations, high speed and desired magnification. Other applications in various fields are security monitoring, retail automation, radiological processing in the medical field and control of guided vehicles.

It is not just the visible spectrum of electromagnetic radiation that is utilised for the inspection and comparison of certain objects like semiconductors, automotive parts, etc. with set standards. Infra-red and X-ray imaging techniques are also used by machine vision systems to detect internal defects.

Main parts of a machine vision system

The essential components of a vision system are:

  1. Image acquiring systems i.e. digital cameras and optical fibres.
  2. Digitizing equipment using a typical frame grabber.
  3. Embedded controller hidden from view, to provide processing power to a camera and reduce complexity of the system.
  4. Image processing software.
  5. Optical or magnetic sensors to be used during counting.
  6. Communication links using protocols like the RS-232.
  7. Actuators in order to pick and place objects.

Working of equipment or machine vision system

The output of the camera is converted to digital format and sent to the computer memory by digitizing equipment. Software will then examine the image after it is manipulated suitably to eliminate stray unwanted signals. A 'fail' indication signals an actuator (robotic arm) to take action against the defective part not meeting the pre- desired standards. If conveyer belts or assembly - line techniques are involved, an alert may be sounded or production stopped till a human operator sets right any problem.