INEX Technologies, an NVIDIA Metropolis partner, designs, develops, and manufactures comprehensive hardware and software solutions for license plate recognition and vehicle identification.
The INEX RoadView solution provides automatic axle counting, vehicle classification, as well as lane zone tracking and triggering using LPR and RoadView cameras. RoadView video-based recognition eliminates the need for costly concrete cutting, in-ground loop maintenance, and axle-counting treadles.
NVIDIA GPUs are used to accelerate the real-time video analysis of the INEX ALPR system, which requires incredibly high accuracy along with high throughput and high frame rates. At the edge, the INEX uses the NVIDIA Jetson Nano and Jetson NX platform and embedded software stack.
Under the hood
INEX video pipeline is based on the NVIDIA DeepStream SDK, which helps achieve super-optimized throughput, and makes it simpler to integrate complex classification and detection algorithms. INEX further leverages some of the world’s most powerful AI productivity tools by integrating NVIDIA pre-trained models and NVIDIA Transfer Learning Toolkit into their development workflow, reducing development time by a stunning 60%. And by going end-to-end with the full stack of NVIDIA hardware and software and deploying on NVIDIA Jetson edge platform, they reduced hardware and setup costs by 60% and lowered operating and maintenance costs by 50%.
The implications and impact for INEX are significant. Leveraging the NVIDIA platform, they can roll out world-class solutions performing challenging real-time vehicle detection and classification read licenses from all 50 states in the US and have expanded to countries in Europe, the Far East, the Middle East, and Australia. Tolling authorities upgrading to the INEX vehicle classification and ALPR system, can supercharge their toll systems quickly and easily – leveraging the latest AI technology.
Read more in our solution showcase.