I need very detailed, about 0.5-1.0 m DEM of the landslide (700 x 400 m). Do it is possible to create such high-quality DEM from the UAV stereo-photos with the semi-professional drone (like DJI Mavic Pro) or LIDAR is much better (and probably much expensive) option for that? I want to try, but is there any with experience in term of the quality of such drone-based DEM models? Are they good enough for more detailed topographic analysis of small-size landscape like a landslide?
LiDAR will probably be more expansive in comparison to the UAV derived point cloud and DEM. USe only drones like phantom -4 and 4 pro. they will yield good results.
And to answer your question, Ortho-mosaics as well as DEM are good for landslide analysis.
Dear Ivica,
At our institute we use the DJI Phantom 4 Pro and DJI Mavic Pro, for creating the DEM surface models. It's relatively easy to do that, but as any process it has some perticularities.
There are different programs available on the market...
https://www.dronezon.com/learn-about-drones-quadcopters/drone-3d-mapping-photogrammetry-software-for-survey-gis-models/
... but mostly AgiSoft (free) or Pix4D (payable) are generally used in the process. The results obtained with that are really good. Compared with the LIDAR models we've even produced better models with UAV. So for more details I suggest you to contact @Rok Ciglič on that matter or via email [email protected].
Best regards,
Jure
Hi, Ivicia!
As already mentioned it is possible to make nice 3D models of landslides with UAV. Phantom 4 Pro has proper camera and you can do very nice models. For example, we made almost 300 photos for an area 250x100 m and produced DEM (3.4 cm/pix). The flight altitude was 30-100 m (the slope of the area is very steep). Than we used Agisoft (also payable!). It is also important to note that the computer must be capable of analysing such amount of data..
If you are asking the price for UAV+lidar the prices are very high - probably not lower than 50.000 eur. DJI's phantom with (normal) camera is around 1500-2000 eur I guess. Of course, more appropriate option is to buy lidar images from a professional company.
Note also: if you have a lot of vegetation in your area, than lidar might be better option. If you make 3d model on the basis of (normal) optical images, you can have problem - high grass, bushes etc. can be recognized as ground. However, some additional work with software can help you with this, but it is hard to say, what's the final output..
Hi Ivica,
If there isn't too much vegetation in your site then drone derived DSMs are totally suitable to what you're looking to do. I also have used a Phantom 4 and have obtained very good results. Typical DSM resolution is
Hi Ivica,
yes, UAV data are of sufficient quality. One of our PhD students has been using UAVs in China for multi-temporal landslide mapping - see Chenxiao Tang). Another former PhD student of mine already showed in a 2013 paper that detailed mapping is possible (link below). One problem can be homogeneous texture of landslide surfaces, as well as accessibility issues to collect ground control points. Standard UAVs such as Phantom 4 have a decent GPS, but it can't match real ground control. Systems with Real-time kinematics (RTK, which for example DJI Matrice 600 has) is better, but you still get positional errors of several decimeters.
Good luck, Norman
Chapter Advanced Image Analysis for Automated Mapping of Landslide S...
Based on my experience fix-wing drones work better than helicopters for large landslides. In my study area a landslide could have a travel distance of 1000 m in elevation and the helicopter drones could not fly high enough. We could not find high enough locations for the drone take off as well.
The cell size of the DEM is usually determined automatically by the software based on your flight height and image quality. We created a 1 m DEM with a DJI phantom 2 with 120 m flight height, several 1 cm DEMs for small landslides with phantom 3 with 20 m flight height, and some 0.3 m DEMs for larger areas with F1000 fix wing drone(Good camera from very high altitude). The critical and essential thing for generating drone photogrammetry based DEMs is the control point. The GCPs have to be well distributed across your study area, and be highly accurate (which means GPS devices on the above mentioned drones are not suitable). In our case we used a RTK device for ground survey.
In my opinion LiDAR is much more reliable as it is possible to create DTMs instead of DSMs in regions fully covered by vegetation.
Dericks Praise Shukla, Astt Professor at IIT Mandi will be useful resource person for you in this discussion. Please directly contact him. He has already done significant work in this area.
Best
Pratik
From your answers and according to some google results it seems like most peple are using DJI Phantom or Mavic drones to create DTM or DEMs- in theory though it should work well with a e.g. Holy Stone HS700 as well, wouldn't it?
- Badges
- Science topic