From measuring nanometers of thin-film to AI inspections of micro defects, our automation team combines a multitude of advanced inspection, assembly, and measurement technologies.
Machine Learning and AI
We have specific expertise in the application of machine learning for vision-based inspections in conjunction with high-precision metrology techniques. Customers benefit from a phased approach of exploration, testing, and prototyping to ensure the final product is manufacturing-ready. Once our machines are up and running, we offer service contracts to maintain and calibrate your equipment on a regular basis.
Using machine learning and artificial intelligence (AI), Lupine Research has addressed an expensive and time-consuming aspect of manufacturing. The automation team combines a multitude of advanced inspection, assembly, and measurement technologies, focusing on the needs of each product to determine the best way to inspect and verify fine visuals. Deep learning algorithms virtually eliminate the most common issues with standard visual inspection, like false calls, fatigue, varied results, and human errors. With high labor, training, and maintenance costs, the benefits of deep machine learning are quickly outweighing traditional inspection practices.
Deep learning algorithms can expedite the inspection process to 45 images per second with quality scores as high as 99.9%, indicating almost a 20% increase compared to just 80% in manual inspection. With deep learning, machine based visual inspection can now be flexible, scalable and easily updated, while collecting qualitative data in areas like texture identification and pattern matching. Statistical results from the algorithms can be compared against human operators for process validation. Together, the inspection process allows for defect identification beyond the surface level, high contrast defects visible to the naked eye.
With the ability to capture precision imaging while in motion, and an extremely fast inline processing time Lupine Research's deep machine learning algorithms are ready for high volume manufacturing and an elevated quality of visual inspections.
Our Approach to Automate Inspection
Trial Case: Glass Insulating Spacers