New Delhi: Test Research, Inc. (TRI), a test and inspection systems provider for the electronics manufacturing industry, has partnered with NVIDIA and other Taiwanese electronics companies to bring AI-driven technologies into smart factories.
TRI will use NVIDIA’s Metropolis for Factories AI technology system in their products to make manufacturing more efficient and better at finding defects to make factories smarter and more productive. TRI also plans to use NVIDIA NIM to boost performance further.
NVIDIA Metropolis for Factories simplifies the creation and management of AI-powered visual inspection tools, making production lines faster, cheaper, and higher quality. Here’s how AI has improved the TRI:
- Better defect detection: Found chip defects with over 99% accuracy.
- Component detection: Improved detection rates for several components, like OSC, MLD, SOD, SOT23, RNET, and CNET, to over 95%.
- X-ray checks: Raised the success rate of X-ray inspections from 85% to 98%.
- Complex components: Reduced mistakes in complex components like the Paladin connector by 88%.
- Reading text: Improved text recognition accuracy from 89% to 99.58% through AI deep learning.
Other AI Tools from TRI:
- TRI AI Smart Programming: Cuts programming time from 60 minutes to 9 minutes.
- TRI AI Verify Host: Uses AI-powered smart repair station to reduce manual checks and costs, working better than human operators.
- TRI AI Station: Uses AI models to improve inspection processes and cut costs using multiple AOIs and smart scheduling.
- TRI AI Training Tool: Helps create AI models with ease for better inspections. It includes tools for classification, segmentation, OCR, detection, continued training, and more.
TRI will also use NVIDIA Triton Inference Server and NVIDIA TensorRT software to speed up AI processes, making it easier for factories to adopt these new technologies.
