Author Topic: Video - LIDAR survey of Cwmystwyth Lefel Fawr  (Read 870 times)

Offline royfellows

  • Moderator
  • forum hero
  • *****
  • Posts: 1466
    • mineexplorer.com
Video - LIDAR survey of Cwmystwyth Lefel Fawr
« on: July 07, 2021, 07:04:05 pm »
This is a video by Ian Cooper of Shropshire caving and Mining Club of a LIDAR survey using home made equipment. Its truly an amazing piece of work.


My avatar is a poor likeness.

Offline Mr Mike

  • menacing presence
  • **
  • Posts: 178
    • The Mine Explorer Society
Re: Video - LIDAR survey of Cwmystwyth Lefel Fawr
« Reply #1 on: July 07, 2021, 07:32:44 pm »
Yes, very impressive Roy. Do you have details of what he used, did he build it from scratch of was it one of the open source scanners / survey devices - irrespective of which a mean feat !

Offline royfellows

  • Moderator
  • forum hero
  • *****
  • Posts: 1466
    • mineexplorer.com
Re: Video - LIDAR survey of Cwmystwyth Lefel Fawr
« Reply #2 on: July 07, 2021, 07:59:46 pm »
The link was sent to me earlier in an email, the first I knew about it. Hopefully, someone may come along who knows more.
My avatar is a poor likeness.

Offline Mr Mike

  • menacing presence
  • **
  • Posts: 178
    • The Mine Explorer Society
Re: Video - LIDAR survey of Cwmystwyth Lefel Fawr
« Reply #3 on: July 07, 2021, 08:24:22 pm »
I wonder if its the Caveatron: http://caveatron.com/  ??

Offline sinker

  • menacing presence
  • **
  • Posts: 184
  • O-Level woodwork BlockbusterVideo Gold Card
Re: Video - LIDAR survey of Cwmystwyth Lefel Fawr
« Reply #4 on: July 07, 2021, 08:40:46 pm »

Not often that I'm speechless...but...WOW!  :clap2:

Ah, well, now, you see...erm...

Offline T pot 2

  • obsessive maniac
  • ***
  • Posts: 477
Re: Video - LIDAR survey of Cwmystwyth Lefel Fawr
« Reply #5 on: July 07, 2021, 08:53:19 pm »
A stunning piece of work.
Fantastic medium which probably took many hours to prepare before publishing,
thank you for sharing this.
T pot

Online ChrisJC

  • Funky
  • forum hero
  • *****
  • Posts: 1568
    • http://www.cowdery.org.uk
Re: Video - LIDAR survey of Cwmystwyth Lefel Fawr
« Reply #6 on: July 07, 2021, 09:00:39 pm »
The mind boggles at the maths involved.

Chris.
--
http://www.cowdery.org.uk
Mines, caves,
Land Rovers

Offline Tomferry

  • stalker
  • ***
  • Posts: 285
Re: Video - LIDAR survey of Cwmystwyth Lefel Fawr
« Reply #7 on: July 07, 2021, 09:30:54 pm »
Am most impressed that has taken some serious work I imagine,  brilliant quality it reminds me of the other topic on uk caving last week lost gold in ww2 that shows you very similar footage of inside the mountains where the tunnels are ,  i am guessing he owns the tools and access to the software to create this ?  Or is this on the most recent I phone and he has done thousands of scans and jig sawd them together and created a brilliant video?

Offline ttxela2

  • menacing presence
  • **
  • Posts: 202
Re: Video - LIDAR survey of Cwmystwyth Lefel Fawr
« Reply #8 on: July 08, 2021, 09:06:58 am »
Amazing stuff, how long did it take to produce?

Offline Rob

  • forum hero
  • *****
  • Posts: 2447
  • The Eldon
    • Eldon PC
Re: Video - LIDAR survey of Cwmystwyth Lefel Fawr
« Reply #9 on: July 08, 2021, 10:22:50 am »
I wonder if its the Caveatron: http://caveatron.com/  ??
I'd heard of this device but never seen any details of it or the open source project. Looks very interesting, would love to hear more about it.  :clap2:
The end is where we start....

Offline RobinGriffiths

  • forum hero
  • *****
  • Posts: 1212
Re: Video - LIDAR survey of Cwmystwyth Lefel Fawr
« Reply #10 on: July 08, 2021, 10:51:19 am »
Blimey. Like the mine truck scene in Temple of Doom, only better.

Offline royfellows

  • Moderator
  • forum hero
  • *****
  • Posts: 1466
    • mineexplorer.com
Re: Video - LIDAR survey of Cwmystwyth Lefel Fawr
« Reply #11 on: July 08, 2021, 01:35:00 pm »
I am going to try and get someone on the forum from SCMC, there are a lot of questions in the air which I cannot answer.
My avatar is a poor likeness.

Offline Cantclimbtom

  • menacing presence
  • **
  • Posts: 203
Re: Video - LIDAR survey of Cwmystwyth Lefel Fawr
« Reply #12 on: July 08, 2021, 01:53:25 pm »
I'm just gob-smacked!  That's truly astonishing.

How many hours/months/years work must that have taken??
Expert in incompetent tomfoolery

Offline royfellows

  • Moderator
  • forum hero
  • *****
  • Posts: 1466
    • mineexplorer.com
Re: Video - LIDAR survey of Cwmystwyth Lefel Fawr
« Reply #13 on: July 08, 2021, 02:00:44 pm »
I know little about LIDAR but to my eyes that video captures the very soul of the mine.
My avatar is a poor likeness.

Offline Mr Mike

  • menacing presence
  • **
  • Posts: 178
    • The Mine Explorer Society
Re: Video - LIDAR survey of Cwmystwyth Lefel Fawr
« Reply #14 on: July 08, 2021, 02:19:03 pm »
Very simply, a laser scans the passage, slice at a time. The time it takes for the laser light to hit a bit of surface is measured, so you get the distance relative to the source (the instrument). This is done 1000's of times on a 360deg rotation. Then you move along a bit, repeat - this starts generating the point cloud, that is then processed in to the fancy graphics.




Offline Ian Adams

  • forum hero
  • *****
  • Posts: 1277
  • UCET
    • UCET Caving Club (North Wales)
Re: Video - LIDAR survey of Cwmystwyth Lefel Fawr
« Reply #15 on: July 08, 2021, 02:23:34 pm »
Absolutely fantastic  :)
A door, once opened, may be stepped through in either direction.

Offline aricooperdavis

  • forum star
  • ****
  • Posts: 568
  • Cornwall to Cumbria
    • Cooper-Davis.net
Re: Video - LIDAR survey of Cwmystwyth Lefel Fawr
« Reply #16 on: July 08, 2021, 02:24:48 pm »
I'm interested in the fancy graphics! How does the rendering make the bright bits? Is it just the point density in the point cloud? Could the points be stiched together into surfaces (non-manually, presumably!)?

Offline Grout1

  • Newbie
  • *
  • Posts: 11
Re: Video - LIDAR survey of Cwmystwyth Lefel Fawr
« Reply #17 on: July 08, 2021, 02:26:05 pm »
I am going to try and get someone on the forum from SCMC, there are a lot of questions in the air which I cannot answer.

Ian Cooper is the main instigator of this project, along with Pete Eggleston and Kelvin Lake - all from Shropshire Caving & Mining Club

Ian is quite happy to provide information and explanations - he just has to work out how to re-register himself back onto UK Caving.
Tricky stuff.

Hopefully you will all be enlightened on the methodology in the near future.

Ironically, Shropshire Mines Trust have been involved with Network Rail's Prometheus project at Snailbeach, where professional LIDAR surveys are being compared with survey data acquired from autonomous drones.

Watch this space - remote mining of the Moon from the comfort of the hot tub is coming next

Offline Mr Mike

  • menacing presence
  • **
  • Posts: 178
    • The Mine Explorer Society
Re: Video - LIDAR survey of Cwmystwyth Lefel Fawr
« Reply #18 on: July 08, 2021, 02:29:16 pm »
I'm interested in the fancy graphics! How does the rendering make the bright bits? Is it just the point density in the point cloud? Could the points be stiched together into surfaces (non-manually, presumably!)?

It is the point density. Closer you are to an object the closer the points, further away they spread out more due to angles etc...

Offline Rob

  • forum hero
  • *****
  • Posts: 2447
  • The Eldon
    • Eldon PC
Re: Video - LIDAR survey of Cwmystwyth Lefel Fawr
« Reply #19 on: July 08, 2021, 05:26:23 pm »
Interesting to see the two different datasets in that video, which suggests there are two different collection types are work.

The yellow (which looks to be a slightly lower data density) is very continuous and used in many of the connecting passages. The white/lighter data is mostly used in the chambers and has clearly visible circles on the floor, presumably below a fixed tripod location.

This mix looks very comparable with the Caveatron system which allows for either a walking passage collection type or a fixed chamber collection type. Either way, i'm very interested to learn more.

...Could the points be stiched together into surfaces (non-manually, presumably!)?
This would be awesome if possible, although my understanding is that most software that do this are not good when applied to cave data. Hopefully that is wrong, or will change soon...
The end is where we start....

Offline IACooper

  • Newbie
  • *
  • Posts: 7
Re: Video - LIDAR survey of Cwmystwyth Lefel Fawr
« Reply #20 on: July 08, 2021, 06:28:58 pm »
Sorry for delays, since being alerted to the thread earlier today I've discovered over the last 15 years since I last posted it would seem my account has been suspended and then someone else has 'grabbed' my original user name  :weep: :weep:

It's nice people seem to like the results I've been able to produce.  I'm happy to try to answer questions and provide more detail. I'll look to try and pick off questions and answer them in seperate posts rather than create one mega-post.

In the mean time here are a few other scans:









Ian
(the original "Mine Explorer")

Offline IACooper

  • Newbie
  • *
  • Posts: 7
Re: Video - LIDAR survey of Cwmystwyth Lefel Fawr
« Reply #21 on: July 08, 2021, 06:37:43 pm »
I'm interested in the fancy graphics! How does the rendering make the bright bits? Is it just the point density in the point cloud? Could the points be stiched together into surfaces (non-manually, presumably!)?

It is the point density. Closer you are to an object the closer the points, further away they spread out more due to angles etc...

Whilst point cloud density will govern how solid the structure looks, this isn't the rendering/shading - that's down to the light shading that's enabled in the software used to process and manipulate the point cloud.   As well as having the X, Y and Z positional data, each point also records the direction of the laser beam that sensed it.  This information is then used by the software to apply shading.

When the viewer is looking along the line of the original laser (ie. you're stood behind the laser looking in the same direction as it is) the point will be brightly illuminated and then the intensity will fade as the viewer moves away from the original laser line.   On this basis if the viewing position is outside of a passage then the 'far' wall will appear brightly illuminated as you're seing what was illuminated by the original laser, whilst the 'near' wall will appear darker because you're seeing the 'shadow' side of the datapoints - you're effectively looking down the line of the laser beam to the source of it rather than seeing what the laser is illuminating.

The result of this is a reasonably effective 3D look.  If I turn the shading off so all points in the cloud have the same level of intensity then everything flattens out and it's very hard to 'see' the 3D structure, especially if you're not moving the point cloud at the time.

Ian.

Offline IACooper

  • Newbie
  • *
  • Posts: 7
Re: Video - LIDAR survey of Cwmystwyth Lefel Fawr
« Reply #22 on: July 08, 2021, 07:23:40 pm »
As a few have suggested - with Rob being most perceptive - the scan is the result of using two different sets of equipment and combining the results.  I left the scan data in two different colours on purpose to show the differences.

The lower resolution 'hoops' of data are indeed the work of a Caveatron.  As I don't have a 3D printer and getting prints done commercially is far from cheap, I built it into an off-the-shelf plastic enclosure.  As impressive and amazing as that project is, it isn't without its quirks and there is a definate technique to getting the best from it (I can go through the procedure used to do the scans seperately if wanted).

In answer to those who ask about putting a 'skin' over the datapoints - yes this can be done and most stuff online you see done by the Caveatron has had this applied.  This is effectively a case of mathematically throwing a sheet over the data.  It does produce a 'solid' surface, but in the process it tends to round all the corners.  It will also look to apply a skin over everything - so if you have a grid it will apply surface over all the holes as well as the bars.   Now if you're looking to scan a structure that generally has fairly rounded surfaces anyway (err, perhaps like a water eroded cave?) or are interested in the larger picture of the whole cave network rather than the finer close up detail then this is all fine.    Personally I'm interested in industrial remains (ie. mines) and I want to go up close and be able to see and recognise the fine detail - I want to be able to instantly see and recognise a 16ft diameter underground waterwheel, not make out there appears to be 'something' in that chamber, but not sure what.  I want to be able to recognise the truck and hopper in Level Fawr, not just see the hint of something a bit different on the one side.    I've found this isn't really the sort of work the Caveatron is suited for.

My answer to this was to develop my own hardware, firmware and software that sacrifies the speed and portability offered by the Caveatron for accuracy and resolution.  The Caveatron is handheld and has to cope with being constantly waved around in real time, relying on accelerometers, a compass and a constant rangefinder distance to a known survey station to be able to work out where it is - and thus where all the data points it measures are.  Unfortunately it doesn't take much to throw off the data - errors in the compass mean if you repeat things twice you get two slightly different versions with an offset between, before you consider metalwork pulling the compass reading.  The biggest problem is the operator - it relies on you moving VERY slowly if you want higher resolutions.  Whilst each hoop of data is pretty high resolution from the rotation of the scan head itself, the distance between each hoop is down to the speed the operator walks/moves forwards.

As Rob suggested from all the portals to another dimension left scattered across the floor - my unit uses the same LIDAR scanner module, but now it is mounted on a stand. After each sweep of data has been captured it will move a stepper motor a very small increment forwards and then capture another sweep of data.  Because it is no longer trying to work out where in the world it is pointing at any given time (with all the errors introduced by that) it means all the data in that captured point cloud is accurately positioned in space and there aren't any funky bits.  By being able to step very small increments it's possible to capture much higher resolution/detail scans - you just have to accept you're in for a wait in each position and you have to carry a stand with you, but given all the surveys have been done as solo trips I have to carry two stands to act as the survey stations for the Caveatron anyway  <shrug>

On the Snailbeach video above, the whole thing has been captured with my stand-alone scanner apart from the section of underground tunnel - which you'll spot has the characteristic 'hoops' of data from the Cavatron - even then the chamber at the end has used my standalone data (at the recent NAMHO conference Kelvin Lake christened it the "CoopaTron" - I'm not so convinced by that! lol).  I also used the straightforward surveying abilities of the C'Tron to survey through the woods from the lower buildings to the upper buildings so they could be positioned correctly in space, I didn't capture any LIDAR data whilst doing that (it would have been a nightmare as each survey leg can only be about 12 yds long or so when using it for LIDAR!)

By using both units I find it possible to combine the strengths of each.  The Cavetron is able to reasonably quickly produce a moderate resolution scan along passageways which would take an eternity and then some for my standalone scanner, but then where there is interesting or recognisable detail I can use the slower more accurate standalone unit to obtain the higher resolution data.


If people are interested I'm happy to go into greater detail on the operating techniques, construction, whatever,  but I'll take a pause for the moment.

Ian.

Offline IACooper

  • Newbie
  • *
  • Posts: 7
Re: Video - LIDAR survey of Cwmystwyth Level Fawr
« Reply #23 on: July 08, 2021, 07:43:37 pm »
Amazing stuff, how long did it take to produce?

I'm just gob-smacked!  That's truly astonishing.

How many hours/months/years work must that have taken??

The Cwmystwyth model was the result of three trips:  The first was about 7 hours non-stop (well, non-stop surveying/scanning - it get's bl**dy boring sat for 3.5 hours watching an automated scan head slowly rotate around before you move it to a new location, set up, hit start and then sit back again!  Lol).  Each of the two subsequent trips were about 3.5 hours each.   After each trip there was then getting on 5 hours or so of processing work back at home.

As part of the Caveatron project is a computer program that will take the raw output from the Caveatron and convert it to XYZ point cloud data correctly positioned according to the underlying Caveatron line survey data.   ...although in practice this data can often require a bit of manual tweaking to get alignments sorted and is why the instructions recommend ensuring you have overlap between the data.

My own standalone scanner saves the data as spherical co-ordinates so I then have to pass that through a computer program I wrote to convert to XYZ data - in the whole model there's one hell of a lot of data points, so need to try to automate as much number crunching as possible!

This is where the fun starts, because I then need to manually align all the standalone point clouds with each other - it's one large 3D jigsaw puzzle because not only can each point cloud be moved in X, Y and Z directions, the cloud can also be rotated around each of those axis as well!  Once those are aligned you can then look to position them in the correct place on the Caveatron data (if combining the two, on both the Snailbeach surveys there was no Caveatron data) - which is where the 'funky' bits from the Caveatron really stand out like sore thumbs.  Finally I can hide the portions of Cavatron data that doesn't want to be seen in favour of the higher resolution scan data.

In theory the Cwmystwyth module could have been done with two 7 hour days underground and then two long afternoons/evenings afterwards.   In practice it was the result of three long Saturdays.


Ian.

Offline Cantclimbtom

  • menacing presence
  • **
  • Posts: 203
Re: Video - LIDAR survey of Cwmystwyth Lefel Fawr
« Reply #24 on: July 08, 2021, 08:46:26 pm »
Is each circular "puddle" a scan location, there's a lot of them, if so that's a lot of 3.5 hours.

How do you tie the scan locations together. Is it a traditional total station approach with a nail in the ground or can you place some kind of reflector in the last location to tie back and chain them as you go?
Expert in incompetent tomfoolery

 

Main Menu

Forum Home Help Search
SimplePortal 2.3.5 © 2008-2012, SimplePortal