Yes, drones are great for multi-temporal assessments of topography which allows for monitoring of displacements.

You can find a large body of literature on applied research on the web as well as here on RG.

Thank you Benni Thiebes. I am looking for some discussable inputs.

Indeed, drones are being increasingly used for landslide reconnaissance/mapping, especially in emergency relief. You may check our work on the 2017 Xinmo landslides where we used a drone to produce a digital surface model "immediately" after the landslide. The case was interesting because it was practically impossible to reach and inspect the mountain ridge where the landslide initiated in person, thus flying the drone really helped us.

Also, we used drones to produce post-earthquake landslide maps: it's convenient because we did not have to wait for the satellite images, so we could obtain a high resolution map very soon after the earthquake. You also can make multitemporal maps by flying the drones again frequently (e.g. after a major rainfall or aftershock to correlate these inputs with landslide remobilization or the occurrence of more landslides).

Yes, drone remote sensing has a wide potential for landslide mapping. The multi temporal images provide you to analyse the pattern of slides. also it needs highly sophisticated image processing technologies. In the case of classification drone images are large scale mapping so it doesn't work with normal classification algorithms. Most probably segmentation algorithms are used for classification.

The critical issue is relative to what are you going to compare the surface mapping of the land slide. Therefore you need to find a stable area on the ground near and around the land slide and establishe a targeted control point field measured with high accuracy. Then process the drone surface mapping using all the time the same control points and you will get the best results.

Gianvito Scaringi Your experience sounds quite interesting. I agree, many a times it is impossible to reach out physically and that is where the drones become a multi-purpose application matter.

As per your recommendation to your 2017 Xinmo landslides, I would definitely like to go through it. Would you please share the link of the same?

Thank you so much for your valuable input.

Sumith Satheendran Yes. I get your point. I find it necessary to have normal data hands on before the landslides. That can make things easier to identify the pattern. The data processing needs some great software and its true that segmentation algorithm is required.

Thank you for your valuable input.

John Hatzopoulos The implementation of what has been shared by you, can be done when the incident happens for the first time at a location. We have certain locations here, in India, where landslide incidents happen every year. There I find the pre-landslide data procuring easy and better method to compare with post-landslide data.

Thank you so much for your valuable input.

It is always good to know some different methods used in different countries for a particular application.

Dear Garvit,

here are the relevant links:

- Article Failure mechanism and kinematics of the deadly June 24th 201...

- Article Some considerations on the use of numerical methods to simul...

- Article Brief communication: Post-seismic landslides, the tough less...

Garvit,

It depends on the required resolution and accuracies to track the motion of the ground.

With a decent high-spec DSLR/lens and a low fly height you can generate a really decent dense point cloud dataset and 3D mesh using photogrammetry techniques. However, it is crucial, especially if performing regular surveys to track movement, that you survey in a spread of ground control points across the site to accurately tie the data into a coordinate system. For this it's generally easiest to use an RTK GNSS system. With a decent camera/survey setup, you can probably achieve sub-cm resolutions at ~2-5 cm absolute xyz accuracies.

If you need greater accuracy to track small strain rates and slow 'creeping' slides, you're best off with a ground-based laser scanner. This can take a long time depending on line-of-site of each scan, but again with a spread of accurate control points, you can get sub-cm accuracies.

David P Harrison I agree with your RTK GNSS point. That is what really required when you are working on something that needs very high accuracy.

Thank you.

Gianvito Scaringi Thank you so much for sharing the links. I will go through it.

Thanks again.

Italians and Americans are working on this pattern. Japanese as well.

Use of drones for landslide assessment would help to achieve the spatial as well as the time-resolution details. The area extent, volume as other many features can be assessed well. I agree with David on his comments.

A team from ITC, Netherlands had studied similar for Gorkha Earthquake 2015 induced landslides in Nepal. I could find something on this in this link:

ftp://ftp.itc.nl/pub/westen/Multi_hazard_risk_course/Powerpoints/Background%20paper%20Spatial%20data%20for%20hazard%20and%20risk%20assessment.pdf

Best wishes for your research !

Here is an interesting video of use of drones in study of landslides in Nepal:

https://www.youtube.com/watch?v=_cZLpBGWOsw&feature=youtu.be

Enjoy learning !

Thank you Badal Pokharel for sharing the research publication link as well as video link of Nepal Earthquake. The video gives some good insights of what has really happened at the spot with real time video. I have also requested Prof. Deepak Chamalagain to share his experience from his Nepal landslide investigations.

I highly recommend this team Pro.Xu (Qiang Xu, Xiujun Dong). They have been devoting to the topic for several years.

I highly recommend my college, Prof. Gonçalo Prates - he's in RG. Get in touch with him.