You are NOT Logged in.
Chat about all aspects of snowsports, backcountry, climbing and mountaineering.
Goto Thread: PreviousNext
Goto: Forum ListMessage ListNew TopicSearchLog In
SnowHwks


Posts: 6
Joined: Feb 2017
Last Visited: 16:57
20th Feb 2017
Using drones for mapping snow depth
Date Posted: 14.00hrs on Tue 14 Feb 17
I'm currently conducting a university research project into the potential of using drones/UAVs equipped with basic cameras to map continuous snow depth across mountain areas in high resolution and on a regular basis, at far lower costs than manual, manned airborne or satellite measurements. The technology is fairly new and consists of stitching multiple images together to a build 3D models of areas when snow-free and snow-covered. By comparing the heights of the detected surfaces it is assumed that changes equate to changes in snowdepth.

The potential applications of this data could have great applicability in ski areas for avalanche monitoring and prediction, especially to check snow distribution following heavy snow fall and winds. In the Scottish resorts it might also be of interest to resort management to locate natural stock piles of deep snow for piste cover. The principles behind the tech have been proved to work in test studies in Switzerland but there remain issues of practicality.

Would this sort of data be of real use to patrol teams? Would patrollers be willing to fly the missions to get the data? What regualatory limitations on drone use would currently prevent uptake? What physical limitations such as wind speeds might prevent use?


I would really like to hear the views of mountain operators and experienced recreationists! Good or bad, discussion is really encouraged on this! Cheers.
sspeirs


Posts: 370
Joined: Mar 2008
Last Visited: 09:59
22nd Apr 2021
Re: Using drones for mapping snow depth
Date Posted: 19.17hrs on Wed 15 Feb 17
OK I am a wee bit of a techie and work in a tech company so I'll preface my comments with appropriate caveats and use my background as a geek as an excuse for my interest here!

First, you should talk to the Scottish Avalanche Information Service (SAIS). I hear they are developing an "avalanche awareness" app (I don't have further details but sounds interesting). It would be good to understand if and how snow depth measurements could relate to risk of avalanche. I know layering of snow has an impact on avalanche risk so if you could detect that, you may have some magic!

On the practicalities of running a drone over mountains, wind on some days could be your biggest challenge. What is the maximum wind speed a drone could operate in?

Also, how much would you rely on communications to communictae your warnings eg. 3G or 4G cellular, Or would you land the drone and download your information from that "offline" as it were. BAck to the 3G/4G, in the UK and more so in Scotland, we are sadly lacking in that area in much of the highlands although in my day job I am part of an initiative looking to solve the 100% geographical cellular coverage problem. And there will be some major improvements announced publicly in the not too distant future from one of the operators. Maybe you could build in your own 4G to your drone .. Lime Microsystems have an open source radio development kit that could maybe help you

You never know, maybe one day climbers will take a drone to shadow them to tell them in advance of potentially high risk avalanche potential areas on the mountain. OK I may be dreaming here but sounds like your idea, if feasible, could help achieve that.

(PS: ping me on sspeirs@cisco.com I can tell you more about Lime Microsystems and their development kit which is being made available to a few universities in Scotland)


Stephen
jabuzzard


Posts: 885
Joined: Jan 2010
Last Visited: 11:02
16th Apr 2021
Re: Using drones for mapping snow depth
Date Posted: 22.17hrs on Wed 15 Feb 17
You will need differential GPS to actually get meaningful depth data, and I doubt your drone will have a suitable GPS receiver.
SnowHwks


Posts: 6
Joined: Feb 2017
Last Visited: 16:57
20th Feb 2017
Re: Using drones for mapping snow depth
Date Posted: 22.53hrs on Wed 15 Feb 17
@sspeirs

Thanks for your reply! I will certainly look to the SAIS for their current approaches. There are ways to detect the grain size at the surface which can indicate different types of snow such as wet, dry powder, even artificial. Exactly how to search further into the snow pack to find layers from the air I'm not quite sure.

Wind is certainly one of the key limitations on many drone systems at the moment. Most systems can only manage stable flight in under 20mph and even below this the batteries drain more quickly as the motors work to stabilise the flight. There are proto-type drones such as the InstantEye from PSI which can remain stable in up to 50mph winds.


At the moment there is no need for cellular connections as the depths need to be processed and calculated on the ground. The idea that the technology might one day be able to operate in real-time is certainly something to aim for!



SnowHwks


Posts: 6
Joined: Feb 2017
Last Visited: 16:57
20th Feb 2017
Re: Using drones for mapping snow depth
Date Posted: 23.03hrs on Wed 15 Feb 17
jabuzzard Wrote:
You will need differential GPS to actually get meaningful depth data, and I doubt your drone will have a suitable GPS receiver.


@jabuzzard

The model is produced using a technique called Structure-from-Motion which does not need highly accurate GPS data to build the model in the first place. I will be using a differential GPS to plot certain ground points which can be seen in the imagery detected by the drone. These can be used to accurately georeference the model produced, and increase the accuracy from the locations given by the drone GPS.
A2


Guest
Re: Using drones for mapping snow depth
Date Posted: 00.17hrs on Thu 16 Feb 17
Hi there

Really interested in this as I have a bit of experience with Photogrammetry/SFM and I'm really curious on what kind of results you are getting given the low contrast of the surfaces and the variable light conditions.

I'm assuming that you are matching your models to fixed points that don't have cover in order to gauge depth?

would love to see more about what you are doing if you'd be up for a chat about it

cheers
edog2009


Posts: 146
Joined: May 2009
Last Visited: 08:15
24th Jul 2017
Re: Using drones for mapping snow depth
Date Posted: 19.22hrs on Thu 16 Feb 17
This is a really interesting idea, if a little ambitious for a student project (under- or postgraduate SnowHwks)?

If something like this could be made to work, it would be extremely useful for the citizen science stuff being done by the snowpatch monitoring crowd, probably a good idea to join their facebook page and see if you can get a discussion about this going there. Iain Cameron has used his drone a lot to make some amazing videos of snowpatches, but your proposal goes a step further I think.

[www.facebook.com]

These papers cover some similar research, using ground-based and differential GPS measurements to measure snowpatch depth and area in the Czech Republic and Slovakia and give some idea of the difficulties that have to be overcome - if this sort of thing could be done in Scotland with a drone that would be a significant piece of work:

[dx.doi.org]
[dx.doi.org]


Eddie Boyle



Edited 1 times. Last edit at 19.24hrs Thu 16 Feb 17 by edog2009.
tim1mw


Posts: 700
Joined: Nov 2006
Re: Using drones for mapping snow depth
Date Posted: 20.41hrs on Fri 17 Feb 17
I know nothing about drones, but here goes.

If a better GPS (or some other piece of kit) were needed, would a typical drone have sufficient carrying capacity for a secondary unit to be added? I'm guessing that the cheaper drones probably don't have any carrying capacity, but that perhaps the more expensive ones would have the ability to carry an optional payload.
SnowHwks


Posts: 6
Joined: Feb 2017
Last Visited: 16:57
20th Feb 2017
Re: Using drones for mapping snow depth
Date Posted: 16.29hrs on Mon 20 Feb 17
A2 Wrote:
Hi there

Really interested in this as I have a bit of experience with Photogrammetry/SFM and I'm really curious on what kind of results you are getting given the low contrast of the surfaces and the variable light conditions.

I'm assuming that you are matching your models to fixed points that don't have cover in order to gauge depth?

would love to see more about what you are doing if you'd be up for a chat about it

cheers


Hi A2,

The issue of homogeneous surfaces is definitely one of the main problems with this technique. Based on some of the work of Bühler at the swiss SLF, one possible solution is the use of near infra-red as snow absorbs more incoming radiation at these wavelengths producing better contrast in images. I have removed the infra-red filter from a small compact camera and am planning to test this first hand using some other filters to cut out the RGB light. Following on from this I might try some lab tests using an infra-red lamp to generate my own light source and control the issues of brightness and variable lighting. This could enable flights to take place at night, when the resort is closed and controls can take place before opening.

Yes, matching the models will use exposed areas such as fence posts and prominent rocks. The other option is to use a UAV with high precision GPS but these systems are currently outside of my budget and might form part of later research.


SnowHwks


Posts: 6
Joined: Feb 2017
Last Visited: 16:57
20th Feb 2017
Re: Using drones for mapping snow depth
Date Posted: 16.42hrs on Mon 20 Feb 17
edog2009 Wrote:
This is a really interesting idea, if a little ambitious for a student project (under- or postgraduate SnowHwks)?

If something like this could be made to work, it would be extremely useful for the citizen science stuff being done by the snowpatch monitoring crowd, probably a good idea to join their facebook page and see if you can get a discussion about this going there. Iain Cameron has used his drone a lot to make some amazing videos of snowpatches, but your proposal goes a step further I think.



These papers cover some similar research, using ground-based and differential GPS measurements to measure snowpatch depth and area in the Czech Republic and Slovakia and give some idea of the difficulties that have to be overcome - if this sort of thing could be done in Scotland with a drone that would be a significant piece of work:





Eddie Boyle



Edited 1 times. Last edit at 19.24hrs Thu 16 Feb 17 by edog2009.






Hi Eddie,

Thanks for you interest. This is part of an undergraduate dissertation so I have to be quite careful to maintain a specific focus to the work, though it would be very easy to spread in many more directions! I am working with the Airborne GeoSciences facility at the university to complete the fieldwork.


I completely agree that the snowpatches would be an ideal application for the technology and would certainly help map their distributions and depths with greater ease. Like I say I have to be careful to maintain a tight focus on my work but snowpatches will certainly now feature in my discussion of wider applications of the technology, so thank you for raising that. I will look to start a thread in that facebook group soon and get the views of the community. Those other two papers are very interesting and I can certainly see areas that drones could quickly improve on.

With regards to citizen science more generally, I think there perhaps needs to be a change in current legislation on drone use, to increase the ease of use. I certainly hope that we will see that in the coming years with efforts to open up drone technologies as compared to the current moves to restrict use.

Thank you very much for raising some excellent points!




SnowHwks


Posts: 6
Joined: Feb 2017
Last Visited: 16:57
20th Feb 2017
Re: Using drones for mapping snow depth
Date Posted: 16.57hrs on Mon 20 Feb 17
tim1mw Wrote:
I know nothing about drones, but here goes.

If a better GPS (or some other piece of kit) were needed, would a typical drone have sufficient carrying capacity for a secondary unit to be added? I'm guessing that the cheaper drones probably don't have any carrying capacity, but that perhaps the more expensive ones would have the ability to carry an optional payload.


Hi Tim1mw,

As I say, with the current set up of using SfM and ground control points there is no real need for complex GPS. The benefits of better precision GPS units would be that the ground control points would not be needed and imagery could automatically be lined up based on calculations involving the precise location of the drone and the camera. This will certainly be of operational use over larger areas and remove the need for access to the ground in remote areas. Such systems include Real Time Kinematic (RTK) GPS units which can deliver incredibly accurate and precise (sub centimetre)locations even during flight. Very few drones systems currently come with RTK GPS fitted but those that do are survey grade and cost many thousands of pounds. Some RTK GPS units are fairly sizeable and might be too big to fit on smaller aricraft, but yes larger platforms can easily take them. Like everything, these systems are rapidly coming down in size and cost, largely driven by the drone market itself, and examples include the Emlid Reach. Again, with in a few years it can be expected that these will start to become the norm!
Your Name: 
Your Email: 
Subject: