Automation technology and machine vision are increasingly merging. Our press agent Peter Stiefenhöfer discussed this exciting development with Peter Keppler, Director of Corporate Sales at STEMMER IMAGING.
This Tech Tip is designed to re-visit the camera interfaces and transmission
standards available for vision systems. Since the last Tech tip on this
subject, over four years ago, there have been a number of new developments.
Embedded Vision has been THE trend topic in the industry for some time now. Rarely in the past has a vision technology been ascribed so much change potential. A large number of exciting possible uses for Embedded Vision systems already exist in virtually all branches of industry and daily life. But will this technology really lead to a complete upheaval in machine vision?
Gigabit-Ethernet for Machine Vision or, in short, GigE Vision: According to
many experts, the new interface standard and the closely associated generic
software interface GenICam (Generic Interface for Cameras) will give new
impetus to the industrial image processing sector in the near future: The
image processing industry finds itself at a crucial technological watershed!
There are many different ways of increasing image processing speed. In the
latest version of Common Vision Blox, STEMMER IMAGING has adopted a new
method: offloading parts of the processing to the PC`s graphics card, which
can boost the speed of some functions by up to a factor of 10.
FPGAs used for data conversion is wide-spread and generally unseen by the user
but when they are brought to the forefront of processing, they have the ability to offload processing power from the CPU and can enable extremely high bandwidths.
The intention of this article is to inform the reader on the definitions, interpretations and implications of the term ‘deep learning’. It remains a hot topic, but terms change and the hope is that this article provides a good basis for further reading.
It deliberately avoids mathematics in favour of diagrams as it is intended only as a top-level description.
Many key tasks in the manufacture of products, including inspection,
orientation, identification, and assembly, require the use of visual
techniques; but implementers need to carefully match machine vision options
with application requirements.
STEMMER IMAGING’s CVB Polimago image recognition tool provides a comparatively simple and low cost solution as an alternative to Deep Learning tools for challenging machine vision search and classification tasks. Fewer training images, shorter training times and faster execution times on a CPU platform combine to avoid some of the drawbacks associated with the convolutional neural network (CNN) approach offered by deep learning. In CVB 2019, Polimago will also be available for embedded applications.