DroneDeploy Ground
DroneDeploy Ground integrationConstruction and equipment rentalReality capture platform

AI inspections for DroneDeploy Ground

Every location on the Level gets looked at after each Walk, not just the ones somebody clicked into, and comes back as a finding with the two dated views behind it.

How Ground capture is structured

Overview

A 360 Walk records the whole floor, which means it records everything you did not go there to look at. Someone walks the route with a 360 camera and the Ground app, the imagery gets pinned to its spot on the drawing, and a week later there is another Walk of the same route. The site is fully documented, and the record grows faster than anyone can open it. RapidEye reads the imagery location by location and returns findings: what changed, where it is on the plan, and the two dated views that show it.

Every location on the LevelThe far end of the floor gets read the same as the unit somebody was already worried about.
Change, not coverageThe same spot on two dates, read against each other, so what moved is what surfaces.
DroneDeploy stays the recordNo capture altered, no Field Note written, no drawing touched. RapidEye reads and reports.

How it works with DroneDeploy Ground

The value of Ground capture is that the route repeats and the imagery is anchored to the plan, so the same place is comparable to itself. The checks are built on that.

  • Walk against Walk. This week's pass and last month's pass at the same location on the same Level, read against each other. A gouge in finished drywall, a fitting that is gone, an opening that used to be covered.
  • Work that was signed off. Trades bill for completed scope. Imagery from the date of the claim is the check on it, and a finding names the location on the drawing rather than the floor.
  • Panos at a fixed spot. Where a Pano is captured at the same standing point each visit, the comparison is close to exact, and small things carry: a stain spreading on a ceiling tile, a crack that was hairline in April.
  • Photos and videos as observations. The single media pins someone drops on a specific condition, with the Field Note attached, are read alongside the surrounding Walk so the observation has context around it.
  • Exterior and interior together. A Level with a drawing and the exterior of the same project belong to one job. Findings from both arrive in one place instead of two review habits.
  • The captures already behind you. Projects that have been walked weekly for a year hold the answer to most disputes. Reading that history back is how a claim gets settled from a dated view instead of from memory.

A finding names the Level and the location on the plan, where in the frame the issue sits, the kind (damage, work not done, a missing item, an unsafe condition), how serious it looks, the two dated views, and a suggested next step. It lands in the RapidEye dashboard, and in a Slack channel or an inbox for the project team if you want it there. Nothing is written back into DroneDeploy: your team decides whether a Field Note or a punch item follows. If your sites are documented some other way as well, phone video, handheld photos, drone imagery, 3D scans, or an inspection format of your own, that is a setup conversation rather than a limit.

How to connect DroneDeploy Ground to RapidEye

  1. Pick the projects and Levels. One project to start, or a portfolio. Each Level needs its drawing uploaded and aligned, which it already will be, since a Level needs at least one floor plan before Walks can be added to it.
  2. Keep walking the same route. Same camera, same cadence, same person if you have one. RapidEye reads each Walk, Pano, Photo and video once it has finished processing, and reads the captures already on the project so each location starts with a history.
  3. Invite RapidEye to the organization. An Organization Admin sends the invite and gives it View only access on the projects it should read. The modal has the details.

How a Ground capture is put together

Checked against DroneDeploy's help centre, September 2026.

DroneDeploy Ground is the on-the-ground half of DroneDeploy, and it is the product the StructionSite team's work became. Everything in it hangs off one arrangement: a project holds Levels, a Level holds an aligned drawing, and every piece of imagery is pinned to a position on that drawing. According to DroneDeploy's help centre documentation (dronedeploy.com), at least one floor plan has to be uploaded to a Level before Walks can be added to it, and drawings can be brought straight in from Procore, Autodesk BIM 360 or PlanGrid.

start end photo pin
Walk path across the LevelLocation on the routePhoto or video pin
The route is the point. A Walk that follows the same path each visit gives you the same locations back on every date, which is what makes two captures comparable at all. DroneDeploy's guidance is to keep a Walk to fifteen or twenty minutes and not to finish where you started, because a looping path causes alignment problems.

There are three ways imagery gets onto the plan. A Walk is a continuous 360 recording made with a supported camera connected to an iOS device over Wi-Fi, started by marking your position with a crosshair and ended by confirming your end point. A Pano is a single 360 frame taken at an exact standing location, with no stitching to do afterwards. A Photo or video is a media pin from the device camera itself, positioned by crosshair, with an optional Field Note attached to describe the condition or tag a teammate. No 360 camera is needed for that last one, which is worth knowing if only one crew has the hardware. Capture works without a signal; the connection is needed to log in and to upload afterwards.

What the platform already does with the imagery

A fair amount, and it is worth knowing before adding anything to it. Side-by-Side Comparison opens two capture dates in a split screen with linked panning and a minimap on each half, so you can move through both at once. X-Ray Mode lines earlier photos up with the live camera feed to see what went in behind a wall or a slab. BIM Compare puts the capture against the model. Progress Reports organise captures by date on the basemap and share by public link, and Photo Reports are curated galleries for a milestone or an inspection. DroneDeploy also sells intelligence add-ons: Safety AI, an Enterprise add-on, scans uploaded Walks and Panos for hazards and saves each risk as a Field Note, and Progress AI, in early access, assigns each trade at each location a status of Not Started, In Progress or Complete.

So what does RapidEye read the imagery for?

Condition, against your standard. The tools above answer where things stand and whether the site is safe; those are their own questions and DroneDeploy answers them well. The question left over is the quality one that a superintendent asks walking a floor: is this finished properly, is that damaged, was that item ever installed, does this match what the last capture showed. RapidEye reads for that, location by location across the Level, on every capture rather than on the ones somebody had time to open, and returns findings against the standard your project cares about rather than a fixed list. The dated views come with the finding, which is what turns a disagreement into a document.

DroneDeploy Ground integration FAQ

Does RapidEye integrate with DroneDeploy Ground?

Yes. RapidEye reads the 360 Walks, Panos, Photos and videos captured with the DroneDeploy Ground app, along with the Level, the position on the drawing and the capture date, and returns a finding for each condition issue it sees: damage, work that was not done, a missing item, or an unsafe condition. Your projects, Levels, drawings and media are not changed.

Does RapidEye need someone to walk the site differently?

No. Whoever does the capture keeps doing it exactly the same way, with the same camera and the same route. RapidEye reads what finishes processing. Because DroneDeploy pins each capture to its spot on the drawing, a repeated route gives the same locations back date after date, which is what makes a change readable.

Can RapidEye read Walks we captured months ago?

Yes, and it is usually the first thing worth doing. On connection RapidEye reads the captures already on the project, so a location has a history behind it rather than a single frame. That history is what lets a finding say what this corner looked like in March and what it looks like now.

How is this different from Side-by-Side Comparison?

Side-by-Side Comparison is the tool for looking, and it is a good one: two dates, linked panning, minimaps on both halves. It still needs a person to choose the location, choose the dates and study the two views. RapidEye does that pass across every location on the Level and only surfaces the places where something changed in a way you would want to know about.

Does RapidEye write Field Notes or change our project?

No. RapidEye does not create, move or edit anything in DroneDeploy. Findings land in the RapidEye dashboard, and in a Slack channel or an inbox for the project team if you want them there. Whether a Field Note or a punch item follows is a decision your team makes.

What if a site is documented with phone video or handheld photos instead of a 360 camera?

That is fine. The DroneDeploy Ground app itself captures Photos and videos from the device camera without a 360 camera attached, and RapidEye reads whatever documentation a team produces, including video walkthroughs and 3D scans, against the standard you set. If your capture looks different from what is described here, tell us what it looks like and we will set it up.

Other construction software RapidEye works with

Platforms where the field team is already producing imagery of the site.

Sources

  • DroneDeploy help centre, Ground capture, Levels, comparison, reports and organization member documentationdronedeploy.com
  • DroneDeploy product pages, DroneDeploy Grounddronedeploy.com

Sources are named at the publisher level and were checked against the publisher's live pages before this page was published.