A road construction firm needs altitude data of road path before and after the construction and the error should be less than 3 cm (2 inch). Normally, they put a ground station and workers manually read altitude data every 10 meters. How this precision would be achieved using drones? The RTK receiver on drone only tells the exact location of drone, not the location (altitude) of the ground below the drone. One way that came to my mind is combining RTK+LiDAR. The RTK gives altitude of drone and laser tells its distance to ground. However, the error in RTK plus error in laser would more likely reach above 3 cm. Thus, how the drone companies solved this issue?

  • What is the measure of altitude error relative to? If there is actually a road, then you probably don't need to fly - your drone could just drive along the road. That could eliminate the lidar part.
    – BradHards
    Commented Sep 6, 2015 at 0:41
  • 2
    I have worked with uav data and using an rtk base station and targets getting much better than 10 cm is difficult. And the distance to the ground is what it solved by the photogrammetry software, depending on different variables the distance calculated from the drone to the ground can be very accurate. The ground control adds the physical location and scale to the project. Be careful of aerial lidar and comparing it to drone data,sometimes it's accuracy is far less and it is a lot lower resolution data. If you are having lidar collected anyway,I would recommend a truck based mobile lidar unit.
    – jtgis
    Commented Sep 21, 2015 at 16:19
  • 2
    Without ground control its all rather meaningless joomag.com/magazine/mag/0625065001446660895?page=30
    – user66841
    Commented Feb 9, 2016 at 8:13

4 Answers 4


Drone companies typically solve this issue by using structure from motion (SFM) software, made available by commercial programs like Pix4D and Agisoft, and open source software like VisualSFM. These software programs create point clouds, orthoimages, DEMs, etc., similar to those created using LiDAR, but with SFM instead. With an RTK mounted on your platform, the accuracy you desire is not unreasonable. It will depend on the altitude of the drone during acquisition, the sensor width, sensor focal length, and image width. You can compute the ground sampling distance (GSD) to determine your required parameters. It will also highly depend on your use and distribution of ground control points for georeferencing, and the complexity of your landscape.

I have achieved this level of accuracy in forested and urban environments, but it sometimes takes a few tries, and adjusting your parameters, to get it right. It sounds like your AOI isn't too complex, so you can probably fly relatively low, and get the results you desire. This project will require some processing software, so I suggest you look into one of the programs I mentioned earlier.

  • Just some quick comments: 1) The technology is called SfM-MVS. SfM finds points in 3D space from overlapping images, MVS densities between these points. 2) Depends on the texture of the surveying area, the z might have a bit higher SD to accommodate your specs. c) There's a wonderful project, about drones and sfm-mvs (github.com/OpenDroneMap/WebODM) which you might want try if you have the aptitude.
    – nickves
    Commented Apr 25, 2017 at 0:54

Be aware of the well-known SfM doming (elevation) error which typically is present when imagery is collected using traditional (linear/parallel) flight lines (drone and/or piloted aircraft). Curved flight lines help mitigate this error. Google "Gently Curved, Convergent, Non-traditional Drone Flight Paths". The justification for curved, convergent flight lines is as follows:

Minimising systematic error surfaces in digital elevation models using oblique convergent imagery Rene Wackrow

Jim H. Chandler First published: 16 March 2011


Results of the simulation process, the laboratory test and the practical test are reported in this paper and demonstrate that an oblique convergent image configuration eradicates the systematic error surfaces which result from inaccurate lens distortion parameters. This approach is significant because by removing the need for an accurate lens model it effectively improves the accuracies of digital surface representations derived using consumer‐grade digital cameras. Carefully selected image configurations could therefore provide new opportunities for improving the quality of photogrammetrically acquired data.


Most UAV image processing software can output a DEM / Point Cloud.

For example ERDAS Imagine will then allow you to do volumetric calculations based on these. The vertical error will depend on the flying height, and resolution of the imagery - and so would differ from project to project.


I've been adapting my curved flight lines for use on long narrow corridors (electric power transmission corridors). For this example, a fixed-wing drone was used flying at a speed of 31mph and a height of 400' AGL. The objective is to mitigate the SfM doming (elevation) error.enter image description here

  • Welcome. How this answers the question? Can you add more details focusing on how the system works and how the precision/accuracy is achieved? Also avoid acronyms first time they are cited. For example, what is SfM doming error? What exaclty did you want to show with that picture? Commented Jan 5, 2019 at 14:27
  • The curved flight lines work equally well for large regularly shaped or irregularly shaped areas.
    – user129783
    Commented Jan 6, 2019 at 18:41

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service and acknowledge you have read our privacy policy.

Not the answer you're looking for? Browse other questions tagged or ask your own question.