Tracking components through supply chains at speed – a case study
The need to track components through supply chains is a vital one. And as these are used in mission-critical systems such as medical devices, autonomous vehicles, or commercial jets this need only increases.
Any point of failure needs to be identified, with suppliers alerted, and systems containing the same part, or from the same batch, found as quickly as possible.
But it’s not a simple task. At the extreme end of the scale would be a complex system with millions of components – a military jet, for example. Each needs to be recorded. And each supplier using a different labeling format. Speed and reliability are therefore issues for any OCR (optical character recognition) system.
Optical character recognition can be used, but to do so at speed is particularly challenging. Here we outline an example developed by Aven Technologies for TMS Electronics, which supplies sectors including aerospace, defense, medical and automotive.
Fig 1: Component reels with multiple labeling systems
How the technology works:
Aven has developed both the hardware and software to not only record but process these labels at speed. Aven has used the SharpVue range of digital microscopes, which gives full HD (1920 x 1080) resolution and up to a 1.4 to 320X magnification (on a 24” monitor).
By using such a system, it is possible to output images at high speed (up to 60 fps) over the USB 3 transmission interface to a PC (running Windows 7, 8 or 10) for OCR analysis.
At the heart of the SharpVue microscope is the Sony FCB EV7520 camera module. This is coupled with two angled LED lights and a tilt table to ensure images are captured perfectly with no glare to ensure its OCR software can work effectively. A display port connection to link to a monitor is incorporated.
Aven has worked closely with Sony for many years, so we were sure that the FCB EV7520 would be able to output the images required from the beginning of the project. The preparation on the actual HD camera and the camera side has been very simple and worked first time for the project. The challenge was, instead, on the software side.
To solve this Aven developed a bespoke software package. This extracts and classifies information on all labels used by TMS Electronics, with an accuracy of ~98% when run at a speed of <4s per label analyzed. All information contained on the label is outputted into a format that can be easily transferred to Excel or an administration system as required. Aven’s software also captures a 1920 x 1080 image for storage if it is needed in the future.
Identifying potential failures before they become a problem is a critical element in system development, and as systems – be it a pacemaker or a car – become ever more complex this need to alert on potential batch failures only increases.
As this case study shows, OCR provides a powerful tool to track parts through the supply chain but while the camera technology – such as Sony’s EV7520 or its 4K modules – is easily capable of supporting such a system the software to extract the data in a non-standard form can prove challenging … but as we have shown, it is not insurmountable.