OMNIQ received U.S. Patent No. 12,633,155 for AI-based vehicle-occupancy detection using infrared imagery and deep learning, with the system paired with license-plate recognition. OMNIQ estimates a representative deployment could generate approximately $1.6 million in annual revenue, within an approximately $218 million annual U.S. addressable market.
Key Highlights
- OMNIQ estimates an annual U.S. addressable market of approximately $218 million for the technology.
- The patent application was filed in 2020 and names Yoram Hofman as inventor.
- Violation rates in diamond lanes have reached as much as 39%, with local estimates averaging 19% to 24%.
OMNIQ Corp. (OTC: OMQS) received U.S. Patent No. 12,633,155 for an AI system that detects vehicle occupancy in moving traffic, extending its technology into HOV lane enforcement.
Patent Uses Infrared Imaging and Deep Learning
The patent application dates to 2020 and identifies Yoram Hofman as the inventor. Its method uses roadside cameras to collect line-scan images in separate infrared bands as vehicles pass a detection line, then applies deep-learning networks to locate windows and validate faces from front and side views. The system combines occupancy counts with license-plate recognition. It also masks occupants’ faces in evidentiary images, consistent with privacy requirements. OMNIQ says the occupancy technology builds on its commercially deployed vehicle-recognition platform and is already active at U.S. and international sites.
HOV Violations Create Enforcement Challenge
Manual HOV enforcement requires officers to count occupants in highway-speed traffic and conduct stops where shoulder space can be limited. The California Highway Patrol has cited violation rates of as much as 39% in diamond lanes during peak commuting periods, while local transportation experts estimate an average range of 19% to 24%. OMNIQ says each violation can reduce travel-time performance for compliant carpoolers and leave toll Revenue uncollected on express lanes.



Please wait processing your request...