Stories – Trends

Trends – How Comet is enabling transformation

Exploiting the full potential of data: Mission launched.

The future of the x-ray business is in digital data. In addition to the x-ray data itself, we have more and more data sources available to us. Our challenge is to effectively use this data to add value for our customers.

The biggest potential lies in the so-called data post-processing. This is where valuable information is extracted for customers from the data captured by x-ray techniques. Only with special skills in data analysis and evaluation can this treasure trove of data be turned into actionable information.


“When I watch our data science experts in Montreal and in Hamburg starting to interact with each other, I know we have high potential as a team.”

Christian Driller

Vice President of R&D, X-Ray Systems

An important strategic step for Comet was therefore the acquisition of the Canadian software company Object Research Systems (ORS) at the end of last year. ORS offers precisely these special post-processing capabilities, thus complementing our existing artificial intelligence expertise.

“The Data Science & Technology team in Hamburg has been in place for several years now. Our strength is in understanding customers’ applications and providing automated defect recognition through evaluation of the image data. ORS’s strength lies in implementing data science solutions, which can be based on machine learning, for example – a great fit in competencies," says Christian Driller, Vice President of R&D, X-Ray Systems.

The existing team at X-Ray Systems has been developing easy-to-use, automated operating controls for x-ray systems for a long time and knows the needs of customers in the automotive, aerospace and electronics industries better than almost anyone else. The data analytics experts at ORS specialize in 3D visualization and image analysis, which is reflected in the success of their flagship Dragonfly product. The fruits of the first collaborative effort will soon emerge in the form of a “guided workflow” in the electronics sector.

“When I watch our data science experts in Montreal and in Hamburg starting to interact with each other, I know we have high potential as a team. But first, it’s time to bring everybody closer together and make it fun to collaborate. Then, the combined innovative minds we have gathered will create and accomplish things that far exceed our expectations.”

Raising productivity: Predicting product characteristics and process faults

“One of our customers’ biggest challenges is to constantly increase their speed of production while avoiding waste. On this front, companies face international competition and must use innovative production processes to continually push the limits of what is possible,” explains Christian Driller.

We can help with this in two areas: First, in predicting product characteristics, by providing our customers with progressively better and more comprehensive information on their products going forward. For example, on how a defective component can be salvaged in the subsequent process steps or how key properties of a piece will turn out. A good example of this is the possibility of sorting batteries into categories based on their performance or their quality grade. Using such sorting, done at a very early stage and without requiring any further testing steps, a battery’s category can determine the subsequent path along the chain of possible production steps and prevent the destruction of value.

“We develop such solutions in close collaboration with our customers.”

The second way we can help is by predicting deviations in the stability of the processes. Soon, not only will product behavior be able to be predicted through data science disciplines, such as machine learning, but the production process itself will also be monitored digitally. Once this stage has been reached, the data generated will inform intelligent machine predictions about the process, with early warnings allowing preventive control measures to be taken.

Crucially, such solutions are not developed in isolation by ourselves, but in the production environment in collaboration with our customers and based on their data. Getting there is likely to keep the joint Montreal and Hamburg team busy for some time yet. But the first and most important step has been taken.