How to detect new construction between annual footprint refreshes

A subdivision breaks ground in April. By June the slabs are poured, by August the frames are up, and by the time your national footprint layer refreshes in November, half the structures that should be in it have been sitting on the ground for months. This is the gap every team running on an annual cadence lives with, and it's worth being plain about rather than pretending it doesn't exist.

Why the gap is structural, not a bug

Vectorizing footprints at national scale from VHR imagery is a real production run: tasking or collecting imagery across the coverage area, processing it, running the structure extraction, and QA'ing the output before it ships. Doing that monthly for an entire country isn't what any provider, FlyPix included, is actually selling. An annual pass is the honest cadence for a base layer that has to cover everything, not a sample.

Construction doesn't wait for a refresh schedule, though. A property-data provider pricing new builds needs an answer within days. A mapping agency fielding a correction request from a local authority needs one just as fast, while the next national pass is still months out. The gap gets closed with a handful of checks run between national passes, not a faster refresh cycle.

What to check in the gap

A few sources close most of the distance until the next pass, and none of them require waiting on imagery:

• Permit and planning records. Most jurisdictions publish building permits or planning approvals, often with a parcel ID and an address. Cross-referencing new permits against your current footprint layer tells you where a build is likely in progress or finished, even before you can see it. This won't tell you the footprint shape, but it flags the parcel.
• Occupancy and utility connection data, where your jurisdiction makes it available, is a stronger late-stage signal than a permit, since it usually only fires once a structure is actually livable.
• Targeted high-res spot checks. If you've got a short list of parcels flagged by permits, pulling a single recent VHR or aerial scene for just those tiles is a different job than a national re-collection. It's manageable for a few hundred sites a quarter even if it's not something you'd run across the whole country.
• Field and service-desk corrections. Local authorities and field teams submitting reports that a footprint is wrong or a parcel has a new structure are a slower but free signal. Routing those reports into a queue reconciled at the next refresh means the next national pass starts from a list of known gaps instead of starting blind.

None of this is continuous monitoring. It's a way to carry a short list of suspected changes forward so the next national pass has somewhere to start instead of a blank slate.

Where the annual layer still does the heavy lifting

The permit cross-references and spot checks are patchwork by design. They cover the structures someone flagged, not the ones nobody noticed: the garage conversion, the ag building nobody filed paperwork for, the parcel where the permit was pulled and the build never happened. A full re-vectorization of the country catches all of that in one pass. That's why it's worth having a reliable annual layer underneath the interim flags, rather than letting the base dataset drift for years between proper updates.

That's the role Building Footprints is built for: a vectorized footprint layer across your chosen area, delivered as GeoJSON or Shapefile and refreshed annually from VHR imagery, so whatever you're tracking between cycles has a solid, current layer to reconcile against rather than a patchwork of open data five years out of date.

If an annual, nationwide footprint layer is the piece missing from your stack, that's what we build.