|
|
Bar code identification: Poor
quality ink jet bar code and characters
This sample application
demonstrates the power of NeuroCheck's integrated concept as
opposed to "dumb" special purpose machines. Of course you
can use a low-price scanner for identifying a bar code, but
as you can see from the image, NeuroCheck can handle bar
codes of quite poor quality without problems. And in
addition to bar code reading it can do a lot of things a
scanner simply is not capable of: perform preprocessing on
the image, compute properties of image objects, send
position parameters to a robot control system and in this
case read characters.
The image shows an electronic
component on a circuit board. Clicking the image loads the
bar code section of the image in full size and full
gray-scale resolution.
Not only is NeuroCheck able
to read the poorly printed bar code without any problems, it
can also use the bar code as a position reference for the
characters. The bar code, seen as a single image object, is
very easy to find. NeuroCheck is able to compute the
position and orientation of this whole group of black lines
with very high precision and use this as a reference for
regions of interest used to identify the printed characters
and digits below the bar code.
The image shows the final
processing stage with the classification results for the
first two characters. As you can see, NeuroCheck identifies
the characters with remarkable certainty. And now compare
the few steps required to solve this application in
NeuroCheck with the programming effort required for
realizing such an application with an image processing
library - and keep in mind that, in addition to the image
processing functionality you see here, NeuroCheck offers a
powerful Windows user-interface, parallel and serial
communication, visualization, standard file format
output... without any extra effort!
|
|
|
 |
