Architecture and Technology for Machine Vision in Industrial Applications with 5G

Typical machine vision applications can be quite demanding both in terms of high availability and bandwidth. However, and this is one of the key learnings from this document, the convergence of vision sensor solutions with new technology like 5G or edge computing can provide solutions which satisfy the communications and control-loop characteristics needed for several use-cases in manufacturing. The ability to offload pre- and post-processing tasks to edge servers or even the cloud in non-real-time applications is a key element to more flexible and cost-effective deployments of machine vision in industrial automation. Finally, the manifold possibilities to reduce the bandwidth of image streams primarily by stan[1]dard compression and encoding techniques may not come into play in applications that must rely on images of high quality. However, in monitoring, tracking, and remote-control use cases, uncompressed video data is of less importance.

Machine vision technology powered by 5G networks will highly depend on the ease of integration. Therefore, it is recommended to make the switch from wired to wireless solutions as easy as possible. The possibility to use the same transport layers based on existing machine vision standards as well as have nearly the same transmission performance available compared to wired interfaces will clearly drive the adoption of 5G in vision-assisted industrial automation and even allow for real-time applications.