01
Upload your photos
Pick the folder of each flight, as many as you have. Only the JPEG photos are kept. A LiDAR cloud is optional, and so is a control point file.
Fully automated processing
Photogrammetry,From Drone Photos to Deliverables
Upload the photos from your drone flight and click Process. Everything after that is automatic, control point marking included. You get a coloured point cloud and an orthophoto, then everything you already use on a LiDAR cloud: classification, terrain models, contours and vectors.
On every plan, Free included • 1 credit per hectare, like LiDAR
Your photo folders
Several flights, nadir and oblique
Several flights, nadir and oblique
Cloud and orthophoto
Then the usual processing
Then the usual processing
Control points
Marked for you, no photo to click
Marked for you, no photo to click
Accuracy report
Measured on independent points
Measured on independent points
Fully automated
No control point to mark by hand
Placing each control point in the photos is the hardest manual step of a photogrammetry job. Here it runs on its own, like the rest of the processing.
Marking by hand
- Open the photos that see each control point and click the target
- Mark the first views yourself before the software follows on the rest
- Automatic detection, where offered, only on the target designs it supports
- Check the marks, then process the block again
Marked by Lidarvisor
- Each point is predicted in every photo that should see it
- A detector finds the mark and an AI agent confirms it
- A mark that disagrees with the reconstruction or your survey is dropped
- The run never stops to ask: reject a mark if you want, you never have to
The whole flow
One upload, one click, one run
Fully automated from the upload to the report, with no photo to open. The photos go through the same project, the same processing panel and the same progress bar as a LiDAR cloud.
02
Read the photo check
Before you process anything, the photos are read: positions, overlap, blur, RTK status and camera calibration. It flags what is worth a look and never stops the run.
03
See where every photo was taken
The cameras appear in 3D over the map, and the photo panel shows every thumbnail, grouped by flight.
04
Control points, marked for you
If you imported control points, Lidarvisor finds each one in the photos and marks it, with no click from you. You can reject any mark but never have to, and Process does not wait: the run finishes the placement itself.
05
Tick and process
Tick Compute photogrammetric point cloud, the orthophoto and any other option, then click Process once. Photogrammetry runs first, then the rest of the run.
06
Collect the results
The point cloud, the orthophoto, the terrain and surface models, everything else you ticked, and an accuracy report on the control points.
Before you process
Your photos, where they were taken
The photo panel
Every photo, one click from its camera
It docks along the bottom of the viewer, like the control point panel.
Each photo appears as a camera in 3D at the position it was taken from. The panel lists the thumbnails by flight, and a click flies the view to that camera or opens the photo at full resolution. Filter by flight, by nadir or oblique, or to the photos that see a given control point, and check each mark where it was placed.
- Unusable photos are set aside with their reason: no position, duplicate, take off or landing
- Leave any photo out of the run yourself, or put a set aside photo back
- Photos stay in your project, count against your storage, and can be deleted like any file
- Photos are never shown through a share link; the cloud and the orthophoto are
Camera positions
Every flight, in 3D over the site
Each photo is drawn as a camera at the position it was taken from, one colour per flight. You see how the flights cover the site and where they overlap before you process anything.
What you get
Deliverables, by what you upload
The terrain and surface models come from the same pipeline as for a LiDAR cloud, so both kinds of project get the same products from the same rules.
| You upload | You tick | You get |
|---|---|---|
| Photos only | Photogrammetric point cloud | A coloured point cloud, then everything you ticked runs on it: classification, terrain and surface models, contours, vectors. |
| Photos only | Orthophoto | An orthophoto assembled from your photos. Without the point cloud it is computed on a simplified surface: right on open ground, buildings and trees may lean. |
| LiDAR and photos | Orthophoto | The orthophoto is built from your photos instead of the cloud colours or satellite imagery, and road, marking, water and rail extraction read it. |
| LiDAR and photos | Photogrammetric point cloud | Your LiDAR cloud is classified as usual. The photogrammetric cloud is delivered beside it as an extra layer, not classified. |
Every run that uses photos also delivers the accuracy report. An orthophoto you upload yourself still takes precedence over a computed one.
Up close
The point cloud, zoomed in
One site, built from the 146 photos of a single flight, then zoomed in as it looks in the viewer.
Vegetation
Every tree crown in 3D
Crowns, gaps and the track along the field edge, coloured from the photos themselves. Classification then separates ground from vegetation, and the terrain model is built from the ground.
Vehicles
Trucks and cars, in their own colours
Two trucks in a yard and a row of parked cars along the road, each one in 3D with the colours of the photos.
Rail and roads
Rails you can follow
The rails and the trees beside them, in 3D. The orthophoto built from the same photos is the one road, marking, water and rail extraction read.
Control points
Placed in the photos, then checked
Import your survey the way you already do, and nobody marks anything. For each point, Lidarvisor predicts where it should appear, finds the mark, and keeps it only if it agrees across the views.
Recognised
An AI agent confirms which candidate in the photo is your mark, on two or three views, and correlation finds the same mark in the other photos.
Measured
The exact position comes from a detector and from correlation across photos, never from the agent's eye.
Filtered
A mark that disagrees with the reconstruction or with your survey is dropped, and a point needs several views to be kept.
Placed
Used in the adjustment, or held back as an independent check point.
Outside the photos
Not seen in any photo. Reported, and not a defect.
Seen but not placed
In the photos but rejected. Reported with its thumbnails, and the run goes on without it.
The deliverable
An accuracy report, not a claim
Every run with photos ends with a report section on the reconstruction and its accuracy. It states how many photos were aligned out of those supplied, how the camera was calibrated, the ground sample distance, and a verdict on the block, with the thresholds stated in ground sample distance so they mean the same on every site.
- The accuracy is measured on check points, control points held back from the adjustment. When you marked none, about one point in four is held back for you.
- A residual on a point used in the adjustment is never presented as accuracy. The report labels it for what it is.
- Every control point is listed as placed, outside the photos, or seen but not placed, with its thumbnails.
- Without control points, the report says so: the absolute accuracy is that of the drone's own positions.
No run is refused or failed on accuracy. A block that did not hold together is delivered with a red verdict that says in plain words what went wrong.
Questions
Frequently asked questions
How much does it cost?
The same as LiDAR: 1 credit per hectare, charged once when you click Process. For a photo project the area is the ground your photos cover, and that one charge covers the photogrammetry and the rest of the run.
Which plans include it?
Every plan, Free included, within the hectares of your plan.
Which drones and cameras work?
Any JPEG photo with a GPS position in its metadata is accepted, nadir or oblique, from one or several flights. The best results come from RTK drones with a FIX position on every photo. When the positions or the camera calibration cannot support a guaranteed accuracy, the photo check and the report say so, and the run still goes ahead.
Do I need ground control points?
No. Without them, the cloud is positioned by the drone's own GPS and the report says that this is its absolute accuracy. With them, they are placed in the photos for you and the accuracy is measured on independent check points.
Do I have to mark the control points in the photos?
No. Lidarvisor finds each point in the photos that should see it, marks it and checks it against the reconstruction and your survey. The marks show in the photo panel before you process, and you can reject one in a click, but the run never waits for you.
Can I mix LiDAR and photos in one project?
Yes. Photogrammetry runs first, so the orthophoto from your photos is ready for the road, marking, water and rail extraction, then the LiDAR processing runs as usual.
Can I add photos after the project is created?
Not yet. Upload every flight when you create the project.
Turn your next photo flight into a survey deliverable
Upload the photos, tick the products you need, and click Process once. The rest is automatic, control points included.
Create free accountPhotogrammetry is included on every plan. Questions? Contact our team or see pricing