Xactools and IDS develop AI-based inspection and handling platform for maker of indexable inserts
June 12, 2024
Xactools, Bönningheim, Germany, and IDS Imaging Development Systems, Obersulm, Germany, have partnered to develop an automated handling and inspection system for a Scandinavia-based global manufacturer of indexable inserts. The DENKnet system utilises artificial intelligence (AI) and high-resolution cameras to enable higher levels of quality control and defect inspection.
Indexable inserts are essential cutting tools used in various industrial applications, particularly in metalworking. Precision manufacturing is required to produce these inserts to ensure exact geometries and surface finishes, as even the smallest defects can be detrimental on the performance or service life of the part. Quality control is an important aspect of manufacturing to ensure the most parts are successfully completed.
In response, Xactools and IDS developed a fully automated and handling system including DENKnet. Through a combination of AI, eight high-resolution industrial cameras and two spider robots for handling, the system can quickly detect defects such as cracks, scratches, inclusion and other anomalies.
“This solution can be customised to any use case and there are no limits, no matter how many ‘classes,’ which camera technology, how large or small the images, or even how mixed the data sets are in terms of resolution and type, for example,’ stated Daniel Routschka, Sales Manager Artificial Intelligence from IDS.
The AI-based image evaluation software, developed by DENKnet, has been trained to recognise colour variations within the powder and differentiate between acceptable colour deviations and actual defects. Additionally, the AI has been trained to identify and classify the numerous geometries and subcategories of the indexable inserts.
One advantage highlighted by Xactools and IDS is the platform’s efficiency. The entire inspection process, including defect detection and picking, takes place in a cycle time of four seconds, achieving almost 100% picking efficiency. The AI image analysis is reported to detect more than 99% of production errors.
“Without AI, the creation of part families and defect detection would be completely unthinkable. With rule-based image processing, the robot would also recognise parts within the standard range as NOK [not OK] and sort them out,” explains Marvin Krebs, Director Technical Sales at Xactools.
Using the Vision AI Hub, DENKnet requires no hard coding, enabling users to embed the software into their own image processing software via an API.
Download PM Review magazine
