Artificial Intelligence is currently an omnipresent topic. The industry is also working on it. The X-ray specialist VisiConsult X-ray Systems & Solutions is one example: The team is currently working on an AI platform with which defects can be detected even more reliably and fully automatically.
X-ray technology has long been used in non-destructive testing. In contrast to optical methods, internal defects such as pores, inclusions, or geometries can also be inspected. Thanks to digitalization, the data is available in real time and can be evaluated automatically. VisiConsult has also been supplying Automated Defect Recognition Systems (ADR) for many years, which detect errors fully automatically after component-specific parameterization. But with production cycles getting shorter and shorter, the demand for faster ramp-up times with ever more complex components and increasing quality requirements; classic image processing methods in X-ray inspection are increasingly reaching their limits.
The VisiConsult specialists predict that the use of artificial intelligence could massively reduce this effort.. Initial tests with available AI frameworks have already achieved detection rates of up to 90% in a blind study compared to conventional methods. The success of an AI depends on the number and quality of the underlying image data. Lennart Schulenburg, authorized signatory at VisiConsult emphasizes, "Data is the oil of Industry 4.0," . Against this background, VisiConsult has developed a cloud solution into which cooperation partners load evaluated image data. These are anonymized and used to parameterize the AI. The higher the quality of the data, the more precise the AI can be trained. The VisiConsult error database now contains several hundred thousand data records with marked and classified errors.
As a specialist for fully automatic inspection systems VisiConsult has now developed a patent for automatic robot tracking: By combining robotics, 3D sensor technology, and image registration, high repeat accuracy can be achieved without expensive component holders. This system is now to be combined with the new AI solutions, reducing the ramp-up process from several days to a few hours. Components are then fed into the system in mixed operation on a conveyor belt and automatically detected by the system. The software would then recognize which component it is and track the robotics to the exact position. This would completely eliminate set-up times and massively increase throughput. As a side effect, the pseudo rejects would be reduced and the detection probability increased. Both are prerequisites for aviation certification of the system.
The North German family-owned company with over 70 employees is currently looking for project partners for the implementation of pilot projects in fully automated testing.