How to Apply Machine Vision to Automate Inspection Tasks White Paper

Application Note | Case Study | Technology Primer | White Paper
How to Apply Machine Vision to
Automate Inspection Tasks
Replacing human vision with machine vision in inspection tasks can
reduce cost and improve product quality. Human vision is unreliable when
used for high-speed, precise, or repetitive inspections. Machine vision
systems have three general capabilities:
1. Location or search finds the position of the object of interest.
2. Identification tells you the type of part.
3. Inspection checks that the object has the proper dimensions, meets
quality standards, is free of specific defects, etc.
First, ask if your task can be done by machine vision. Vision tasks that
require a lot of “real world” knowledge or have parts that have large
variations can be difficult to automate. For example, we find cracks in
a textured, glass part using our knowledge of what cracks look like and
moving the part so it is “just right” to see cracks. This is not easy for a
machine vision system. Machine vision is best used when the parts have
known shapes and where defects have limited variation and are clearly
Next, ask what type of vision system is appropriate. Some simple
inspection tasks can be done by a “vision sensor” that does one kind
of measurement, such as measuring the depth of a hole in the part.
Most machine vision tasks require a machine vision system with these
- Motion control to bring the part into the camera’s field of view
- A sensor to detect when a part is in the camera’s field of view
- Special lighting and a way to reduce stray light
- A machine vision camera and lens
- A vision processor – a computer specialized for machine vision
- Vision processing and control software
Figure 1 - Diagram of a pill inspection system
The vision processor has to be programmed for the vision task. Until
recently, this was often a difficult and time-consuming task. DALSA IPD’s
focus is on easy-to-use machine vision software that can quickly solve
your vision inspection task. For example, in our iNspect™ software, vision
operations are presented in familiar terms, such as “calipers”, and an
intuitive user interface guides you through the set-up of your vision task,
without requiring programming. Figure 2 shows the iNspect user interface
for the blister pack inspection system.
DALSA IPD provides cameras, vision processors and vision software,
and recommends other components. Our partners include automation
distributors who supply and integrate all the components into machine
vision systems.
Figure 1 shows how these components are combined to make a machine
vision system that checks for broken or missing pills in a blister pack
card. A conveyer belt moves a card into the camera’s field of view.
Diffuse lighting reduces highlights from the plastic covering the pills. A
vision sensor detects when the card is in position and triggers a VA30
Vision Appliance™ to take an image of the card and process it to detect
broken or missing pills. The VA30 is one of DALSA IPD’s compact vision
The VA30 and vision software locates each pill and identifies defective pills
by inspecting the pills’ areas, perimeters and intensities. Inspection and
identification results are displayed for monitoring by an operator. When
a card with defective pills is found, the VA30 sends a signal to a rejection
mechanism that removes the card from the product stream.
Figure 2 - iNspect’s operator’s screen when doing blister pack inspection
With DALSA IPD’s cameras, vision processors, software and, of course,
support, you can quickly apply machine vision to automate inspection
700 Technology Park Drive, Billerica, MA 01821
USA Tel: +1 978.670.2002 Fax: +1 978.670.2010
Email: [email protected]
DALSA is an international leader in digital imaging and semiconductors and has its corporate offices in Waterloo, Ontario, Canada.
All trademarks are registered by their respective companies. DALSA reserves the right to make changes at any time without notice. ® DALSA 2010. 031010_wp_mv_inspection