Machine Vision Cameras
Are there different types of Machine Vision cameras?
Broadly speaking machine vision cameras can be divided into two types:-
-
If your
object is moving along a conveyor it will need to be either a progressive area
scan camera or a line scan camera. Basically for fast moving events (faster
than the eye can see) there are two types of camera
-
· Progressive Area Scan- this type of camera has the ability to read
the image as a whole (rather than an interlaced camera that reads
two distinct fields (odd and even lines) separated by 40ms time interval
and then the resultant image is read out as a complete frame). Where
on fast moving objects the interlaced camera gives image blur (because
of the time difference the two fields are read out by the image sensor).
The progressive cameras read all lines within the same scan and therefore
no image blur is visible.
-
· - Line Scan cameras. Sometimes area scan cameras do not have the
speed to capture data from a moving object. (example paper or textiles
which may travel at many tens of meters a second) These applications
demand cameras which can read a line of data very fast. Also normally
in this application the web is very wide sometimes many meters so
therefore a high resolution camera is required. To deal with these issues
a linescan camera is needed. Line scan cameras are a linear image sensor
(generally one row of pixels in the sensor – up to about 8000 pixels).
Linescan cameras read data at many thousands of lines per second so
can deal with defect detection in very fast moving objects.
Should I choose a Monochrome or Colour camera?
Generally our advice here is if you don’t need colour detail always choose monochrome.
There are two reasons
- 1) Colour image sensors are monochrome sensors with a matrix colour filter across them.
There are a number of different filters used but all filters will degrade the image sensor
sensitivity by around 30 per cent. That means you will have to compensate with more
light or lower iris setting to let more light into the optic.
- 2) The other problem is that in single chip colour cameras the resolution of the colour
is degraded. This is because the filter colour for one pixel will be different from its
neighbours. By software correction the resolution is corrected but it’s a assumed
correction so it will not be as accurate colour representation as that derived from
a three sensor colour camera where there is a sensor for each of the prime colours
(blue, green and red) and these are converged to overlap one another by clever
optical mirrors in the camera head. For really true colour representation a three
chip colour camera has no equal.
What’s the difference between CMOS and CCD?
This is a big subject in itself but we will try and be clear on the main differences
- CCD is more sensitive than CMOS mainly because the CCD chips generally have
100% fill factor where the CMOS is much less ( this means the CCD is 100%
active sensor while the active part of the CMOS will be no more than 70%
and some a lot less.
- CCD is much better for low contrast images. This is because of the lower inherent
noise in the sensor.
- CMOS has the advantage of being much more flexible than CCD. You can window
CMOS sensors to read out less data at a higher frame rate (i.e.: although a sensor
may have a resolution of say 1280 x 1024 and readout rate of 15 frames per
second by windowing the sensor and only reading out a 640 x 480 portion of the
image you can achieve a frame rate of nearly 70 frames per second could be achieved.
- CMOS sensors have much lower power consumption and therefore are ideal for
portable devices or space applications.
What type of output should my camera have?
The type of output will be determined by how you want to read the data out from the camera.
- Fire Wire (IEEE1394) offers a solution that is also “plug and play” and some cameras
are now available with high resolution.
Camera Link- the digital interface standard for those demanding applications. Is used
with many high performance digital area scan and also line scan cameras. You will
need a frame grabber designed for that type of interface. The price of this interface
is decreasing so is not as expensive as you may think.
-
Gig-E – Gigabit Ethernet interface is relatively new but it will grow in popularity because it allows huge amounts of data to be transferred quickly and over long distances
There are also older types of LVDS RS644 interfacing still used for digital cameras but these
are being phased out and replaced by Camera Link or Fire wire so is not recommended for
new developments.
How do I choose a lens for my camera?
A camera is of no value without a lens or a focusing optic.
For most applications the lens choice depends on:-
- Field of view required
- Working distance ( distance between front face of lens and object being viewed)
- Size of detail required
- Depth of field required
- How you are going to use the data captured. (Accurate measurement applications
for example may require the use of a telecentric or machine vision quality lens)
- Features required from lens (most lenses are manual iris and focus but other options
are available such as zoom and motorised lens control.
This guide is not extensive and is no substitute for speaking to one of experienced sales engineers who will be able to assist in the selection of the right camera and lens combination for your application. Also visit the lens section of our web site. We have now included an index selection guide that helps you select the right manufacturers for the type of lens you are looking for. Or please call our technical sales desk on 01635 30345.
If you wish to download a copy of this article, click here...
To return to the FAQs & Applications Main Menu, click here...
Further information available from Geoff Smith.
Alrad Imaging, a trading division of Alrad Instruments Ltd.
Alder House, Turnpike Road Industrial Estate Newbury, Berkshire RG14 2NS
Telephone: +44(0) 1635 30345 Fax No. +44(0) 1635 32630
Email: sales@alrad.co.uk
