CS2 Color Scanner
Delta Computer Systems, Inc., 11719 NE 95th Street, Vancouver, Washington 98682-2444 (360) 254-8688 10
email@deltacompsys.com http://www.deltacompsys.com
Color Tolerances
You want a bunch of paperweights in the shape of our logo.
You sketch the thing, add dimensions, and send it off to a
turn-key manufacturing shop with these instructions:
Dimensions: ±0.035"
Weight 1 LB, ±1 oz.
Color: Red; just like the sample.
To keep costs down, you've given generous tolerances on
dimensions and weights, but what is your color tolerance?
The manufacturer might think he's got an excellent match,
and you might think it clashes terribly. It gets worse. If you
are matching items that are not colored the same way, such
as plastic and metal, you have to take into account the
lighting as well.
Most colors are made up of a mixture of pigments and can
look very different under varied lighting. Different
pigments are used for paints, plastics, inks, dyes, etc.
Perfect matches of different materials under varied lighting
conditions is virtually impossible.
So how do you express color tolerances? Color can be
expressed in mathematical terms; intensity, hue, etc., and
can have mathematical tolerances, but this will seldom
produce optimum results.
Why is that? In order to keep manufacturing costs down,
tolerances should be as wide as possible. But color
tolerances are based on perception, and are not evenly
distributed. In most cases, larger variations in intensity are
permissible than in hue.
Tolerances in hue are often not even. Butter, for example
can be quite orange without cause for concern, but even
slightly green butter is likely to be repulsive.
Creating a multidimensional tolerance color model is
possible, but would take a lot of time. Even if you went to
this expense, on line, high speed testing against this model
would be very expensive.
In short, color is a perception issue, and is still more of an
art than a science. Fortunately, in many real-world
situations, a one dimension test for color is all that is
required; green areas on carrots, white ends on asparagus,
dark knots on wood, paint either there or not there, etc.
These types of applications are suited to the CS2 scanner.
Pencil Sorting Example
In this hypothetical application, red pencils are being
examined for missing erasers and missing paint. The CS2 is
used with an optimizing computer in order to make length as
well as color decisions on the product.
The pencils travel lengthwise, erasers first, under the CS2
which is configured for a viewing area of 1/32" by 1/2". A
rotary pulse generator (RPG) provides a pulse for every
1/32" of pencil travel so that the entire top side of pencil is
examined. The CS2 is triggered from this RPG.
The optimizing computer reads the raw analog color
information, subtracts the average background intensity, and
determines when a pencil is in the viewing area -- Base color
amplitude 15% or more than average Base amplitude, pencil
length -- counts RPG pulses while Base amplitude is high,
eraser present -- counts RPG pulses while Sense/Base ratio
is 110% or more than average Sense/Base ratio at the
beginning of the pencil, paint present -- checks that
Sense/Base ratio does not go high again during remainder of
the pencil length.
The system was set up, adjusted and it ran great. Except,
after maintenance shutdowns. Sometimes it would just
ignore the pencils.
A little troubleshooting pinpointed the problem. During
maintenance, the machine was cleaned using air nozzles.
The fine sawdust settled everywhere, including on the pencil
conveying belt. The sawdust was light colored and highly
reflective, making the background intensity readings as high
as the pencils!
A simple continuous belt cleaning system took care of the
problem.
