Using machine vision for automatic detction of a tracer dye
Working with a client new to machine vision, WyeTec developed imaging software and assisted with hardware design, optics and lighting for a dye-detection project.
Providing a positive IT experience:
Our client’s project required automatic detection of a tracer dye, and measurement of percentage coverage of the dye on a sample. They approached
WyeTec having had a less-than-successful experience with a third party software developer, who had failed to realise the project specifications.
Having been previously let down, it was important to gain the trust of our client, and to demonstrate our competence and ability to deliver. It was
agreed that a small-scale prototype would be developed to prove the camera integration and detection facets of the project; and if this was
successful, the project would proceed to developing production-ready software.
WyeTec demonstrated the initial prototype within 3 weeks, much to the relief of our client, and the go-ahead was given for the production version
of the software. This extended and solidified the prototype, featuring multiple camera support and tools for detection of the sample, the tracer,
and calculation of percentage coverage.
Finding the right cost-performance balance:
The application was targeted at a high-volume, cost-sensitive product. To keep works costs down, we elected to build the application on the
open-source imaging library OpenCV, for which a developer licence is required, but no run-time licence.
The application was targeted at a low-cost single board computer (SBC) with built-in Arduino GPIO functionality that could be used for
issuing trigger signals. Low-cost but high-performance CMOS cameras were sourced, with a manufacturer agreement to supply them