Over the course of 2021, SICK Automation launched a set of deep learning software and services called dStudio, making Artificial Intelligence (AI) more accessible and economical to the southern African market. This software works with machine vision solutions and simplify quality inspection of components, assemblies, barcodes, surfaces, food produce, logistics and more. This offering is especially useful in operations that have defied automation in the past and were only distinguishable through human inspection.

SICK’s deep learning software allows users to quicky and directly set up AI image classification onboard SICK smart devices. This allows these devices to use specially-optimised neural networks to make decisions automatically and run accurate and reliable inspections.

Released in March 2021, dStudio is a SICK web service that can be used to train neural networks that are optimised for various SICK devices, simply by inserting sample images of correct/incorrect examples.  The image inference is carried out directly on the device in a short and predictable decision time, without the need for an additional PC, where results are sent to the control as sensor values. Unlike its predecessors, dStudio removes the laborious process of developing rules and algorithms to identify patterns or defects. This removes the time-consuming process for harder-to-identify patterns/defects, such as leather creases, different nuts or wood grain.

dStudio offers an intuitive, step-by-step user interface, making it accessible to users without skilled AI knowledge. For more experienced users, the SICK AppSpace software platform allows users to create and customise their own deep learning sensor apps. Training progress and success are shown in clear graphics so that the trained neural network can be assessed prior to running it in an operation. Further saving time and costs of implementation, all system training takes place in the Cloud, removing the need for additional hardware or software training.

SICK deep learning is supported by the Inspector P 621 2D vision sensor, and the SIM 1012 programmable Sensor Integration Machine with SICK’s Picocam or Midicam streaming cameras. In time, SICK deep learning will also be enabled across both SICK smart 2D and 3D vision sensors, as well as SICK data processing gateways.

“We believe that SICK’s deep learning products are the future of automation. Being accessible and economical, this technology can be easily implemented across Africa, offering massive boosts in efficiency, productivity as well as safety,” says Grant Joyce, Head of Sales, Marketing and Product Management at SICK Automation South Africa. “We are very excited to see this rolled-out into the South-African market” he concludes.