State DOT Validation Shows Reliable Traffic Data at Scale

Transportation agencies need traffic data they can stand behind.

From safety investments to pavement management and travel demand modeling, transportation decisions rely on accurate traffic counts and a clear understanding of roadway use.

To validate performance in real-world conditions, Flock partnered with seven State Departments of Transportation (DOTs) to evaluate traffic data accuracy through side-by-side comparisons on active roadways.

Across more than 2.3 million vehicles and 3,600+ hours of testing, Flock cameras delivered strong, reliable results across varied geographies, roadway types, and operating conditions.

The Challenge

Transportation agencies need accurate traffic data without adding operational complexity, long deployment timelines, or unsafe roadside collection methods.

Traditional traffic counting approaches can require lane closures, field crews, or temporary equipment deployments that take time to install and maintain. Agencies also need confidence that modern computer vision systems can meet the accuracy expectations required for planning and operational decision-making.

The question was straightforward:

Can computer vision-based traffic monitoring deliver reliable, FHWA-aligned traffic counts and vehicle classifications in real-world conditions?

The Approach

Flock partnered with transportation agencies across Arizona, California, Connecticut, North Carolina, Ohio, Tennessee, and Texas to validate system performance across multiple roadway functional classifications in both urban and rural environments.

Validation included:

• 2.3M+ vehicles analyzed
• 3,600+ hours of testing
• 25 devices deployed across 7 states
• Multiple roadway types and traffic conditions

Testing combined two validation methods:

Third-Party Video Annotation

Independent reviewers analyzed roadway footage frame by frame and compared it directly against Flock detections.

Results included:

• 95% video-verified volume accuracy
• 85% video-verified classification accuracy

Side-by-Side Traffic Count Comparisons

Flock traffic data was compared against existing count systems over longer durations.

Results included:

• 94% overall volume accuracy
• 87% overall classification accuracy

The Results

95% Volume Accuracy Across 2.3 Million Vehicles

Across millions of observed vehicles, Flock cameras consistently delivered reliable traffic volume counts in live roadway environments.

88% FHWA Classification Accuracy

Vehicle classification performance remained consistent across multiple roadway environments and classification levels.

Validation Across Diverse Conditions

Testing spanned:

• Urban and rural deployments
• Multiple roadway functional classifications
• Varied geographies and operating environments
• Long-duration live roadway testing

The result was a broader understanding of how computer vision-based monitoring performs in real transportation environments — not controlled lab conditions.

Built for Safer, Smarter Traffic Monitoring

Flock’s validation efforts demonstrate that computer vision-based traffic monitoring can deliver reliable, FHWA-aligned vehicle classifications and counts while simplifying deployment and reducing field complexity.

The platform is designed to help agencies:

• Collect accurate traffic data safely
• Reduce operational lift in the field
• Deploy more quickly across roadway environments
• Integrate data into existing transportation workflows and systems

With accountability, usability, and operational flexibility built in, transportation agencies can modernize traffic monitoring without compromising confidence in the data.