During the manufacture of an AimValley telecoms product (the Rover), each product is tested/verified on several elements. Items such as a barcode, torque indication bolts, presence of thermo pads, rotation Elcos, presence of kit, scratches, labels, etc. are checked.
We have already started with the automation of this process. Photographs are taken from the inside and outside of the product. These photographs form the basis of the automatic control system.
The current scripting to check the accuracy of the photographs is slow and incomplete. Next to this, there are more automatic checks available in the market than is available in our current scripting.
These photographs can be analyzed through image processing algorithms and even Artificial Intelligence can be part of the options.
The ultimate goal is to implement a set-up that can check the products that are in manufacturing phase, within one minute. Application of Atificial Intelligence and self-learning is preferred.
The complexity of this project lies within the ability to understand how manufacturing errors can be detected/recognized in the photographs as well as how to automate this.