OMNIQ wins US patent for AI occupancy detection in HOV enforcement
OMNIQ Corp., an OTCQB-listed machine vision company based in Salt Lake City, has been granted US Patent No. 12,633,155 B1 for a system that automatically determines the number of occupants inside a moving vehicle. The patent, originally filed in 2020 and naming engineer Yoram Hofman as inventor, covers methods that pair dual-band infrared line-scan cameras with deep-learning networks to count faces through vehicle windows at highway speed.
The technique exploits the differential infrared reflectance of human skin relative to vehicle interiors, seats and clothing. Roadside cameras capture two separate infrared bands as a vehicle crosses a detection line; comparing the bands sharpens contrast around occupants' faces. A deep-learning model then locates windows, identifies faces within them, and cross-validates the count across front and side views. The system links the occupancy count to licence plate recognition data and, the company says, blurs faces in evidentiary images to meet privacy requirements. The technology is already in live operation at two US sites and one international deployment.
Market sizing and commercial context
OMNIQ estimates that a representative single-site deployment of the patented technology could generate approximately $1.6 million in annual revenue. The company puts the total US addressable market at roughly $218 million annually, derived from approximately 136 HOV and high-occupancy toll facilities nationwide, after excluding jurisdictions where automated occupancy enforcement is legally restricted. The Federal Highway Administration has counted 145 such facilities across the country, providing an existing infrastructure base.
Those figures are management estimates rather than independently validated numbers, and investors should read them alongside the company's forward-looking statement disclosures. OMNIQ is traded on the OTCQB market, which carries lighter reporting obligations than the major exchanges, and the company has not disclosed committed contracts or revenue backlog tied to the patent.
Chief executive Shai Lustgarten said agencies had been "asking for a reliable way to enforce occupancy requirements without putting officers on the shoulder of a freeway," and that the company intends to keep investing in intellectual property that differentiates its platform.
Competitive and regulatory landscape
Automated HOV enforcement sits at the intersection of computer vision, tolling infrastructure and transport law. Manual enforcement is costly and inconsistent: the patent's own background section cites California Highway Patrol data showing violation rates reaching 39 percent of vehicles in a diamond lane during peak hours, with transport experts estimating a 19–24 percent average. That compliance gap represents both a congestion and a revenue problem for tolling agencies.
The competitive field for roadside AI imaging includes established tolling-technology vendors and a growing number of computer-vision startups. OMNIQ's differentiation rests on the dual-band infrared method, now protected by the patent, combined with its existing licence plate recognition stack deployed across thousands of US sites.
From a regulatory standpoint, automated occupancy detection is already legally restricted in some US jurisdictions, which the company acknowledges by excluding them from its market-size calculation. Any broader rollout will need to navigate state-level vehicle-code provisions and, where facial detection is involved even transiently, biometric-data statutes that several US states have enacted in recent years. OMNIQ's stated policy of blurring faces in stored images is consistent with minimisation principles under frameworks such as the Illinois Biometric Information Privacy Act, though individual state rules vary.
The patent grant strengthens OMNIQ's intellectual property position ahead of what could be a competitive procurement cycle as transport agencies seek to reduce manual enforcement costs. Near-term milestones to watch include named agency contract wins, published accuracy benchmarks from the live deployments, and any expansion of the technology into adjacent markets the company already serves, including airport kerbside management and controlled-access facilities.