"We are seeing that the industry is becoming increasingly open to AI-based image processing," explains Carsten Strampe, Managing Director of Imago Technologies. "This is mainly thanks to easier handling." The range of applications is expanding enormously as a result: "AI solutions are a natural choice for applications that are difficult to solve with previous machine vision methods. In terms of tasks, the user is increasingly in control of the images that an AI system is to learn. For example, they themselves define error classes, IO or NOK images, without being dependent on the image processing experts."

Accordingly, Strampe's company presented the 'Vision Cam AI.go' industrial AI camera at SPS, which is designed to provide users with an easy introduction to deep learning. "The camera is easy to operate, even without prior knowledge of deep learning and industrial image processing," explains Carsten Strampe. "The camera also scores highly in terms of data security. Sensitive error images from production or from the machine remain in the camera; cloud access is not required."

Carsten Strampe, Imago Technologies: "We are seeing that the industry is becoming increasingly open to AI-based image processing."

© Imago Technologies

The Vision Cam AI.go measures 50 mm × 50 mm × 40.1 mm (W × D × H, without tube) and offers a color or monochrome CMOS image sensor with global shutter and a resolution of 5 MP (2560 × 1936 pixels). The image processing system has been developed primarily for end users who have little or no experience in AI, programming or image processing. The camera is particularly suitable for inspecting and sorting test objects that exhibit a high degree of variance. It is designed to classify objects into two to five so-called classes. Without any programming effort and supported by a web GUI, users can teach in their own company-specific distinctions in the form of different inspection objects or production deviations by uploading a few images for each class. The Vision Cam AI.go then learns new images independently. Within a few minutes, the camera is ready for use as a fully functional inspection system. It is connected to the outside world via digital interfaces such as 1 Gbit/s Ethernet and digital I/O. Users also have full control over their image data. The teach-in process and image storage take place directly in the camera - and the images can be deleted there if required.

"With the Vision Cam AI.go entry-level camera, Imago Technologies is removing the barriers to using AI so that even companies without expertise in deep learning can benefit from the latest technology," summarizes Carsten Strampe.