We are actually re-mapping the image on our camera. With the right optics and lens we are putting the image directly on the CCD sensor. We are actually charging up each one of the pixels for the amount of light that we have mapped at the part. This will provide contrast.
Cameras only see the light reflected from the objects toward them. Keep in mind that machine vision is all optically based.
There are three elements affecting proper lighting in a machine vision application
The lens is one of the areas that many people overlook when evaluating lighting applications. For example: If you are using real small field-of-views (high magnification) with extension tubes with a high F# we are losing light.
No! The lower the physical F# on the lens, the better its ability to collect light.
We have two sources of light available:
Think about a gauging application in which light passes through a lens or prism in trying to make a measurement of 2 lines. If I am using white light to do that, as the white light passes through the lens and tries focusing on an element, there is a certain dispersion angle because the lights different wavelengths are going to slow down at different speeds and will not converge at the exact same point. If we use one wavelength of light (Monochromatic -LED) products it will always focus at the same point.
Many times ambient light in the environment will affect the machine vision results.
Long pass filters allows only wavelengths from a certain point up to pass through the lens. You can easily mount it with the filter thread of your lens. It screws right on the front of your lens and allows only certain wavelengths to pass through.
http://www.machinevisionfilters.com
If you completely shut off your light source and you still see your part, you are getting a large amount of ambient light which must be controlled.