Book a call
Drone // Label type 20

Construction site labeling from drone imagery

Construction site labeling marks what is on a building site in drone survey imagery: structures by construction stage, foundations, earthworks and stockpiles, machinery and site cabins. Repeated on every flight, it shows how the site progresses and where material is.

Also calledstockpile annotation · construction site annotation · earthworks monitoring

Straight-down survey image of a large construction site in Cologne: a steel hall frame, concrete buildings under construction, foundations, earth mounds and finished blocks.
RawLabeled
  • Under construction7 polygons
  • Foundation1 polygon
  • Stockpile3 polygons
  • Site cabin4 boxes
  • Finished building9 polygons
Example: Clouth quarter, Cologne, Germany, spring 2016, aerial survey image (Land NRW, Bezirksregierung Köln). Licence: Datenlizenz Deutschland – Namensnennung 2.0. Labels drawn by IKT Spīdolas for illustration.

What construction & stockpiles labeling is

Contractors and developers fly their sites weekly or monthly. Labeling the same classes on every flight turns the imagery into a progress record: which building reached which stage, how stockpiles grow and shrink, where machinery and materials stand. Stockpile volumes need elevation data, so piles are outlined on the orthomosaic and measured against the drone surface model. The example is a 10 cm survey image of the Clouth quarter in Cologne during construction in 2016.

How we label it

Classes and construction stages are agreed with you first (for example foundation, shell, roof, finished). Each structure is outlined and given its stage; piles, machinery and cabins get their own classes. Later flights are labeled against the same IDs, so changes can be followed object by object.

Who uses construction & stockpiles labels

  • Progress tracking

    Structures labeled by stage on every flight, for reports to owners and lenders.

  • Stockpiles and earthworks

    Pile outlines for volume calculation from the surface model, flight by flight.

  • Site analytics models

    Machinery, materials and structures labeled to train detection models for site monitoring.

  • Permits and safety

    Site cabins, fences and access routes mapped against the approved site plan.

How a project runs

Four steps, each ending in something you can check before the next one starts. The drone image annotation covers formats, quality checks and cost in more detail.

  1. Brief

    Area, dates and class list agreed in writing, with example images for every class and its edge cases.

    You getWritten labeling guideline

  2. Imagery

    Your satellite or drone imagery, or satellite scenes sourced for you: free Sentinel and Landsat, or commercial high-resolution.

    You getScene list with dates and resolution

  3. Labeling

    Annotators work to the guideline. Anything the imagery cannot settle is flagged, not guessed.

    You getLabels plus a log of flagged cases

  4. Check & delivery

    Every class is reviewed against a sample before the data leaves us.

    You getData and per-class accuracy, in your format and CRS

Questions about construction & stockpiles

Can you calculate stockpile volumes?

Volumes come from the drone surface model. We outline each pile and its base, so the calculation covers only the pile.

How often should a site be flown?

It depends on the project; weekly or monthly flights are common. Each flight is labeled against the same classes and IDs.

Can you use imagery from our own drone?

Yes. Orthomosaics from the usual processing software, for example DJI Terra, Pix4D or Agisoft Metashape, can be labeled directly.

Related label types

Next step

Need construction & stockpiles labeled?

Tell us your area, your imagery and what your model has to learn. We reply with a proposed class list and a quote.

Book a call