Buildings: hero banner

EarthDefine Building Footprints 3D

193M+ properties mapped, with heights and elevations. Available off-the-shelf.

EarthDefine Building Footprints 3D gives you GIS-ready building polygons for the entire United States: 193 million buildings mapped from high-resolution imagery and LiDAR, with rooftop outlines over 98% accurate. Every building is tied to a street address, and 167 million carry heights, estimated stories, and volume. Available off-the-shelf and refreshed as new imagery and LiDAR become available.

193M+
properties mapped
>98%
rooftop accuracy
167M+
structures with 3D heights
Clutter Boulder: imagery basemap
High-resolution aerial imagery (Boulder, CO)
Clutter Boulder: buildings over imagery
AI-derived building footprints, on the imagery

Why EarthDefine building footprints

Accurate, current, and ready to drop into your workflows.

  • Rapid updates. Always current on imagery and LiDAR. Our AI processes millions of buildings monthly, so you are never working with outdated data.
  • Accuracy. Make confident decisions with detailed roof outlines at over 98% accuracy.
  • Addresses. Every building is connected to a street address for seamless integration with your workflows.
  • 3D data. Access building heights, estimated stories, and volume for 167M+ structures.

167 million structures, modeled in 3D

Heights, estimated stories, and volume for RF simulation, line-of-sight analysis, and full 3D city models.

How industry leaders use building footprints

Building Footprints 3D provides GIS-ready building data to support a host of mapping and spatial-analysis functions.

  • Broadband mapping: accurately map and report fixed and mobile broadband coverage with unprecedented granularity.
  • Improved geocoding: go beyond street-level or parcel-centroid geocoding to accurate building-rooftop locations.
  • Enhanced mobile marketing: target customers based on their proximity to a building geofence.
  • Insurance underwriting: use distance-to-building locations to model property risk from floods, landslides, fire, utility lines, trees, and more.
  • 5G planning: create accurate models with 3D buildings for RF simulation.
  • City planning: understand urban growth, map heat islands, conduct line-of-sight analysis, and generate 3D building models.
  • Change analysis: map buildings across years for detailed insight into residential and commercial construction over time.
  • Building monitoring: find unapproved structures by comparing building-permit data against ground structures.
  • Population mapping: use dasymetric mapping with building footprints to visualise demographic data precisely across regions.

See your area in the free building samples

Browse free samples

Data coverage and freshness

We refresh nationwide building footprints on a rolling two-year cycle, updating about half the states each year. As our AI and post-processing advance, we make quarterly improvements. Every dataset reflects the latest publicly available imagery and LiDAR.

Shared: EarthDefine coverage map (US, 2026)

See our building data in action

Explore Dallas, TX in 3D: building footprints with heights, estimated stories, and volume, draped over 2020 aerial imagery.

Frequently Asked Questions

What is the source of the building data?

The majority of buildings are classified from USDA NAIP (National Agriculture Imagery Program) imagery, supplemented by publicly available LiDAR data where available.

What is the quality and resolution of the imagery?

The NAIP imagery is a mix of 30-60cm spatial resolution, georectified by the USDA to +/- 6m at a 95% confidence level (CE95).

How are the buildings derived?

Buildings are classified from imagery using state-of-the-art deep-learning algorithms. Once classified, they are converted to vector format, regularized (squared up), split by parcel boundaries, addressed, and merged statewide.

What is the accuracy of the buildings?

Buildings are over 98% accurate on a pixel-by-pixel basis, validated against dozens of urban areas across the country. By Intersection over Union (IoU), the accuracy is over 94%.

What is the update frequency of the building product?

Because NAIP imagery is flown state-by-state, buildings are updated whenever new imagery becomes available. Most states fund NAIP every other year, so a nationwide refresh occurs roughly every two years. We also update buildings when we significantly improve our AI or post-classification processing; updates ship quarterly, with the most significant changes arriving alongside new imagery or LiDAR.

Why do some geographies have older data?

Beyond lags in available NAIP imagery, some areas retain older data where NAIP is not available or has not been updated. Military bases and other sensitive areas are examples where recent NAIP is not authorised for public use. Occasionally NAIP imagery does not meet EarthDefine's quality standards, in which case older imagery may be used to derive higher-quality footprints.

Do you track when a building was first seen or last updated?

Not yet. We do not currently track when new buildings appear or how they change over time. The date associated with a building refers to the year of the imagery it was mapped from, not when the building appeared or changed.

Related Products

Social card: Impervious Map

Impervious Map

Buildings, roads, and paved surfaces as a 60cm impervious raster.

Learn more
Social card: Parking Lots

Parking Lots

Vector polygons of US paved parking, down to 1,000 sq ft.

Learn more
Land Cover juxtapose: Grand Junction CO land cover (750)

Land Cover

Multi-class land cover: 7 ground-cover classes at 60cm.

Learn more

Ready to upgrade your building intelligence?

Talk to our team about a nationwide or regional Building Footprints 3D license, or download a free sample to evaluate the data yourself.

Speak to a Data Expert Download a free sample