Moving AOI beyond the grey world
of many articles—so we shan’t re-visit it memory or storage
here. than the same im-
An empirical method of defining the age, with the same
acceptable window of tolerance for any physical dimensions,
given component or camera field of view, rendered in 8-bit
employed in most AOI systems in one digital greyscale.
form or another, is to test a number of More on storage and
‘good’ boards to define the outer limits processing solutions
of acceptance. But capturing images to later.
determine acceptability presupposes that A 16-bit grey-
you have a discerning imaging system in scale image gives the
the first place. And this is not always so. AOI processor more
Among the key differentiators between data to work with Figure 2. Melf diode in greyscale
modern AOI systems are the deployment than an 8-bit image. (top) and 24-bit colour (bottom).
of digital cameras with all the associated This has nothing to
arguments about resolutions and field of do with the resolu-
Figure 3. LEDs in greyscale (top) and
view (FOV), of lighting techniques, of im- tion of the camera, which simply relates 24-bit colour (bottom).
age processing technologies, and of colour to the amount of pixel data to process; in-
depth. It’s the latter that we’re discussing stead it is about the depth of data available
enced human operator. And as with all
primarily here. for processing. With more shades of grey to
effective automation solutions, can do it a
The great majority of AOI systems represent subtle variations in colour on the
lot faster and continuously.
capture colour digital images from the inspection subject, the process can be more
24-bit colour demands more processing
CCD or CMOS camera sensor but use discerning—a good start to achieve better
power and greater data storage capacity.
grey scale processing for data analysis. Sure, performance indicator values. But it’s only
Today however, cost-effective memory and
almost all display a live image in colour on half a step in the right direction.
hard drive storage options are abundant.
the equipment monitor to allow the opera- To really get the maximum depth of
And Moore’s Law still holds true in the
tor to see what’s going on—but they only data from a captured digital image—and
exponential growth of computer processing
process greyscale, either as 8-bit or 16-bit therefore to have the most data with which
power over time.
images. to discern subtle characteristics of the
All AOI systems from Marantz Busi-
Greyscale is a range of shades of grey imaged subject—full colour processing is
ness Electronics use 24-bit colour technol-
without apparent colour. The darkest required. Full RGB colour requires that
ogy with several colour transformation
possible shade is black, which is the total the intensities of three colour components
models and deploy every graphic designer’s
absence of transmitted or reflected light. be specified for each and every pixel. It is
favourite tool to cope with the 24-bit im-
The lightest possible shade is white—the common for the intensity of each compo-
age processing workload—an Apple Mac
total transmission or reflection of light at nent (R, G and B) to be stored as an 8-bit
at the heart of every machine. Macs are
all visible wavelengths. Intermediate shades integer, and so each pixel requires 24 bits
renowned for their exceptional graphics
of grey are represented by equal brightness to completely and accurately specify its
capability.
levels of the three primary colours—red, colour. Better systems are also transform-
An obvious advantage of 24-bit colour
green and blue (RGB) in camera systems. ing this raw 24 bits RGB data to different
AOI is the ability to discern colours for
The brightness levels of the red (R), colour models with the ability to switch
certain components and printed legends.
green (G) and blue (B) components between the different models. By doing so,
Combined with synthetic modelling prin-
are each represented as a number from new parameters are being created such as
ciples that start with a ‘real’ picture, very
decimal 0 to 255. Every pixel in an RGB hue (the pure colour) and saturation (the
accurate yet flexible models can be quickly
greyscale image has the value (R + G + colour strength) alongside the brightness
produced that incorporate an optimum
B)/3. The lightness of the grey is directly (greyscale). With these parameters, apply-
tolerance, which alleviates the possibility of
proportional to the number represent- ing tolerance values work more reliable as
an operator wrongly ‘adding’ an incorrect
ing the brightness levels of the primary colours are spread logically in the model
component to the acceptability range—
colours. Because there are 8 bits in the and close to reality. Image formats that
thereby widening the window so much as
binary representation of the grey level (256 store a full 24 bits to describe the colour
to make the inspection meaningless.
= 28), this imaging method is called 8-bit of each and every pixel are known as 24-bit
Detecting colour on components
greyscale. colour images.
is useful for melf resistor banding and
In some systems that use the RGB Using 24 bits to encode colour infor-
orientation of melf diodes. It’s also very
colour model, there are 216, or 65,636, mation allows 224 or 16,777,216 (16.8
useful for checking LED colours. Greyscale
possible levels for each primary colour. million approx.) different colours to be
struggles with the difference between red
When greyscale is being calculated in this represented. By having more information
and green—the most common colours for
system with (R + G + B)/3, the image is per pixel, there is the freedom to calcu-
LEDs. But component colour checking
known as 16-bit greyscale 65,536 = 216 late with specific characteristics of that
is only the tip of the iceberg. And like an
which is the binary representation. As information that is just important for a
iceberg, it’s the depth that matters.
with 8-bit greyscale, the lightness of the specific test point; this means simply more
Among the common causes of false
grey is directly proportional to the number possibilities and greater flexibility. So an
alarms are simple factors like board colour,
representing the brightness levels of the AOI imaging system that captures and
board surface finish and the contrast
primary colours. As you’d expect, a 16-bit processes 24-bit colour should at least be
digital greyscale image consumes more capable of discerning as much as an experi- Continued on page 29
www.globalsmt.net Global SMT & Packaging – August 2009 – 27
Page 1  |  Page 2  |  Page 3  |  Page 4  |  Page 5  |  Page 6  |  Page 7  |  Page 8  |  Page 9  |  Page 10  |  Page 11  |  Page 12  |  Page 13  |  Page 14  |  Page 15  |  Page 16  |  Page 17  |  Page 18  |  Page 19  |  Page 20  |  Page 21  |  Page 22  |  Page 23  |  Page 24  |  Page 25  |  Page 26  |  Page 27  |  Page 28  |  Page 29  |  Page 30  |  Page 31  |  Page 32  |  Page 33  |  Page 34  |  Page 35  |  Page 36  |  Page 37  |  Page 38  |  Page 39  |  Page 40  |  Page 41  |  Page 42  |  Page 43  |  Page 44  |  Page 45  |  Page 46  |  Page 47  |  Page 48  |  Page 49  |  Page 50  |  Page 51  |  Page 52