Process automation in minutes

This can help companies save costs: beh.digital GmbH, a spin-off of L3S, has developed a learning camera that can automate industrial processes and quality control quickly and easily.

Conventional image processing systems, such as those used in quality assurance, have to be developed for specific applications and are time-consuming. The costs are correspondingly high. Artificial intelligence can solve the problem, but companies often lack the necessary know-how – especially medium-sized companies. The smart cameras from beh.digital show that this can also be done quickly and easily. Using machine learning methods based on camera images, they can automate industrial inspection processes in a fast-track process. Instead of classic image processing, the customer takes pictures of the process and then categorizes them. A few minutes after the artificial neural network of the camera system has been trained with the image data, it can already distinguish objects or detect defects.

The image processing algorithms were developed at L3S and the Institute for Microelectronic Systems at Leibniz Universität Hannover – originally for autonomous driving vehicles. However, they also show first-class recognition performance in other industries. The camera’s software reduces the need for knowledge, so that even small and medium-sized companies can smoothly automate complex recognition tasks with just a few clicks. A unique selling point of the solution: the training of the algorithms and the data processing take place entirely within the company. As a result, there are no data protection problems. Another plus: The software and the highly energy-efficient hardware are optimally matched.

beh.digital GmbH was founded in 2018 by doctoral students at Leibniz University Hannover and has won several awards. Experiences from numerous customer-specific developments during the doctoral period have been incorporated into the smart cameras.