Many companies use red LEDs for illumination. Their benefit is cost, cycleability
and operation life. Red LEDs do a good job of illumination for gray scale
cameras and gray scale image processors. The red works well for reflective
surfaces like solder and masks out other non-reflective elements and features.
However, single color LEDs work like looking at a photographic negative image,
only the bright areas appear. In order to see other features the lighting angle
needs to be changed or adjusted to optimize the image that needs to be
analyzed. This is where the LEDs ability to be cycled on and off quickly
compensates for the need to change illumination angles quickly. Several
images, each using a different lighting angles, can be captured and then
analyzed. However, this can cause inaccuracies between each of the images
due to camera movement or table movement.
Only a select few manufacturers use multicolor lighting. This method provides
many advantages. Red, green and blue are the three basic elements of color.
Every color can be defined using the correct combination of these three colors.
When the three basic color elements are used to illuminate a non-reflective
object, the three colors combine and the object reflects it’s true color back to the
camera. That is why the human eye can differentiate between the vast variety of
colors. Wouldn’t this color concept also make sense to for the function of image
analysis and inspection? Seeing in color would allow color a image processor to
differentiate between the wide variety of colors on a PCB such as the substrate
color, the component color and the silk screen color. Gray scale systems can
often have difficulty in differentiating between a solder bridge and silk screen line
resulting in false call and escaped defects.
The second reason color lighting is effective for PCB inspection is the ability to
image the solder fillet. It is accepted that camera imaging is a 2 dimensional
technology. Only when we employ cameras mounted at multiple angles do we
start to gain some dimensional data about an image. However, this technique
requires multiple cameras, mounted at multiple angles in order to correlate 3
dimensional information from a 2 dimensional image. This process,
unfortunately, increases the complexity, cost and single point of failure in a
imaging system. (Image Tri Color Lighting)
As we know, a topographic map image uses different colors to represent different
elevations in the terrain of land. This technique can also be employed using
color light for the inspection of solder fillets. If red, green and blue light are
shown onto a reflective surface from different angles, each light will reflect
independently off independent angles of the reflective surface. This techniques
works well for solder fillets. Using this technique, a color, 2 dimensional image of
a solder fillet can now provide 3 dimensional data about the shape of a solder
fillet. This allows an image processor (or a human eye) to determine the shape
of the solder fillet and determine it’s quality and integrity.
Image Processing
Image processing technology is the last in three key elements to inspection.
Frame grabbers provide the capability of handling and processing the images
captured by the camera. Off the shelf frame grabbers offer a low cost
commercially available product for AOI OEMs. An additional benefit is that they
are relatively easy to integrate into an inspection system. However,
commercially available systems lack the ability to be customized to the
application specific requirements of printed circuit board inspection. This in turn
limits the capabilities of inspection delivered to the customer.
Custom frame grabbers or customized image processors are the next hardware
alternative for image analysis. They have a potentially higher cost but provide
much more flexibility in providing a customized solution for PCB inspection.
Specialized algorithms can be developed for the task of inspecting PCBs. The
customer will then receive a product that is tailored to the application of PCB
inspection. The software interface can also be designed to use terms that are
related to the PCB industry such as “shifting”, “fillet”, and “wettability”.
Image processors either process color or gray scale images. Gray scale image
processors are less expensive, however they lack the amount of image detail
that can be captured due to their limitation of 256 shades of gray. However, a
color image processor has the ability to process true color images without a
reduction in detail. It is a waste of money to use a system that captures a color
image, maybe uses color lighting but then analyzes the image in gray scale. This
process defeats the purpose of collecting color image data. Color image analysis
provides the maximum amount of data available for image analysis. An added
benefit is that, for simple inspections, color systems can imitate gray scale
systems by using only a brightness level. This means that a color system can
provide the benefits of both gray scale and color image analysis. (Images Omron
Color Lighting, PCB Neutral Lighting Gray, PCB Neutral Lighting)
In Conclusion
There are a wide number of technologies available to do PCB inspection. When
combined properly the customer will receive a robust system that is easy to use
and performs the task of inspection accurately with consistent results. The only
way to achieve this accuracy is to combine the three best elements of inspection.
The combination of multi-color lighting, three element color cameras, and true
color image processing will provide an inspection machine built for the accurate
inspection of both component and solder faults on assembled printed circuit
boards.
|
