Using machine vision for high-speed, high-accuracy product counting
An interesting combination of both a Machine Vision application and migration of a legacy application, WyeTec were commissioned to do a
"faithful" translation of software originally coded using Borland C++ Builder 6 to a modern .Net / C# application.
Future-proofing legacy applications:
The project remit was to keep as close as possible to the original capture and analysis routines, though it was understood that some
changes would be required as certain legacy components could no longer be supported. Some further changes were identified as desirable
to help future-proof the application and ensure good programming practice.
Windows Presentation Foundation (WPF) was chosen as a development platform to provide a rich user interface, targeted at touchscreen
devices running Windows IoT (effectively Windows 10 embedded).
Embracing new technologies:
The original application used manufacturer-specific libraries to capture image data from a cameralink line-scan camera that was
approaching end of life. The routines were replaced with manufacturer-agnostic code that would support any suitable gigE Vision
compliant camera, with hooks put in place to allow for expansion of the capture and analysis routines in the future.
The work included moving system parameters that had been hard-coded to settings in an SQL database, improvements to
system responsiveness and significant improvements to the range of samples that could be handled by the machine.
Phase 2 of the project has seen development of new analysis methods and additions to the range of supported hardware, significantly
expanding the range of materials that can be handled by the machine.