Computer vision for smarter streets

See traffic.
Understand motion.
Act sooner.

Sense Traffic Pulse turns existing camera feeds into real-time intelligence—revealing how every road user moves, where risk emerges, and when action matters.

LIVE ANALYSISIntersection 04 · Las Vegas
42
31
38
FLOW RATE1,284 veh/hr
AVG. SPEED34.7 mph
ANOMALIES0 clear
SCROLL TO DISCOVER
ONE INTELLIGENCE LAYER

From pixels to
decisions.

Five purpose-built capabilities work as one—from raw video and coherent motion regions to autonomous mobility, crowd intelligence, and operational alerts.

01

Intersection Flow Engine

Transforms live roadway video into structured movement intelligence—tracking direction, speed, density, and traffic flow across complex intersections.

OPTICAL FLOWEDGE AIREAL-TIME
Explore repository
02

Anomaly Detector

Surfaces the moments that matter: wrong-way movement, stalled vehicles, density spikes, fragmented flow, and unexpected region shifts.

FASTAPIALERTSJSON API
Explore repository
03

CMR Pack

The coherent-motion foundation. Builds, clusters, and tracks motion regions so crowded scenes become clear, actionable patterns.

PLA-2049MOTION REGIONSVISION
Explore repository
04

Physical AI · AV

A Python toolkit for multi-sensor autonomous-vehicle research, including video, egomotion, transforms, interpolation, and dataset interfaces.

MULTI-SENSOREGOMOTIONAV DATA
Explore repository
05

Crowd Density

Real-time CSRNet and OpenVINO density mapping with spatial zones, temporal aggregation, annotated output, and four alert modes.

OPENVINOCSRNETZONE ALERTS
Explore repository
COMMAND CENTER

One view.
Every signal.

The Sense Traffic Pulse™ dashboard brings feeds, processing pipelines, system models, frame analysis, and autonomous tracking into one operational view.

  • 10,228 connected feeds
  • 8 processing pipelines
  • 6 tracking assets
Visit Sense Traffic Pulse™ ↗
Sense Traffic Pulse dashboard showing connected feeds, vision processors, models and live tracking
VISION STACKDETECT · SEGMENT · TRACK
Crowd density estimation visualization
PEOPLE FLOW INTELLIGENCE

Measure density.
Recognize change.

Density-map regression handles dense and occluded scenes where individual detection becomes unreliable. Zone-level temporal analysis identifies sustained occupancy, sudden spikes, rapid drops, and accumulation trends.

SUSTAINEDSPIKEDROPTREND
THE VISION IN MOTION

Intelligence that
moves with your city.

See how computer vision becomes a continuous pulse of safety and operational insight.

HOW IT WORKS
01

Ingest

Connect existing traffic camera feeds.

02

Interpret

Map coherent regions and movement.

03

Detect

Identify risk, conflict, and anomalies.

04

Act

Deliver alerts, metrics, and APIs.