I’ve noticed several reasonably high level events in the past year or so where maps have been resurveyed using LIDAR data resulting in new maps that seem to me to be a backward step from earlier versions, particular with regards to contours. Contours seem to meander across the page, including erroneous detail and missing significant features, and generally not showing the shape of the ground. Anyone else noticed this, and if so any ideas why?
I’ve not used a LIDAR base map but I understand you get reasonably accurate 1 metre interval contours. I’ve read how some mappers produce a base map by taking every fifth contour, which clearly isn’t going to produce an good picture of the ground – missing out detail shown only on intermediate contours, but I wouldn’t have thought that’s any worse than photogrammetric plots or OS base maps.
Is it that LIDAR is considered more accurate so people don’t think the contours need adjusting and form lines adding to show ground shapes? Or perhaps less experienced mappers feel more confident having a go at the contours themselves?
Do most mappers produce base maps by taking every fifth LIDAR contour? Does OCAD automatically produce base maps for you? I would imagine a lot of work could be done prior to field survey - producing a base map by tracing over the 1m contours, moving contours and adding form lines to pick out the smaller features and pick out the breaks of slope more accurately. Perhaps professional mappers could offer base map production from LIDAR data as a service for amateur mappers?
LIDAR Base Maps
Moderators: [nope] cartel, team nopesport
18 posts
• Page 1 of 2 • 1, 2
LIDAR Base Maps
To oblivion and beyond....
-
buzz - addict
- Posts: 1198
- Joined: Mon May 09, 2005 10:45 pm
- Location: Sheffield
Re: LIDAR Base Maps
Lidar base maps are just that - "base maps", they need a lot of work to produce a top quality orienteering map, arguably just as much as using a pg plot. The contours on Lidar are incredibly precise (to within a few cm) but they are not great at showing what the ground looks like to a person moving through it (ie an orienteer). They need to be pushed and pulled to emphasise features that are clear to the eye and to reduce those that are less apparent. Producing a map by simply selecting Lidar contours and doing nothing with them "because they are correct" is a very foolish approach and will result in maps being poorer than previous versions (even if dimensionally exactly precise.
- Big Jon
- guru
- Posts: 1895
- Joined: Tue Nov 14, 2006 11:59 am
- Location: Dess
Re: LIDAR Base Maps
There is a nice example of this at one of the control sites that we're using for the TWT next weekend, on the side of a small ring-contour hill. The controller and I both agreed that the contours on the (pre-lidar) map didn't really match how either of us perceived the shape of the land on the ground. Overlaying the 1m lidar contours made it very easy to see what the problem was and (crucially) made it very quick for us to make a minor tweak to the map to align more closely with reality.
However, the lidar contours had completely missed the adjacent dot knoll, which was already on the map and was patently obvious to anyone standing next to the hill. Presumably the interval between lidar readings was such that it had failed to take pick up the highest point of the knoll and whatever rise had been detected had been interpolated out when DSM was generated.
If we'd had blind faith in the lidar, we'd have omitted the dot knoll entirely. So, as Big John says, lidar is definitely a base - and a very good one - but it certainly doesn't remove the need for interpretation by the mapper.
However, the lidar contours had completely missed the adjacent dot knoll, which was already on the map and was patently obvious to anyone standing next to the hill. Presumably the interval between lidar readings was such that it had failed to take pick up the highest point of the knoll and whatever rise had been detected had been interpolated out when DSM was generated.
If we'd had blind faith in the lidar, we'd have omitted the dot knoll entirely. So, as Big John says, lidar is definitely a base - and a very good one - but it certainly doesn't remove the need for interpretation by the mapper.
British Orienteering Director | Opinions expressed on here are entirely my own, and do not represent the views of British Orienteering.
"If only you were younger and better..."
"If only you were younger and better..."
-
Scott - god
- Posts: 2384
- Joined: Wed Jan 17, 2007 4:43 am
- Location: in the queue for the ice-cream van
Re: LIDAR Base Maps
Also, I just spotted this very interesting Facebook discussion on this very topic: https://www.facebook.com/photo/?fbid=10163361721226982&set=gm.6759488550825582&idorvanity=485564718218028
British Orienteering Director | Opinions expressed on here are entirely my own, and do not represent the views of British Orienteering.
"If only you were younger and better..."
"If only you were younger and better..."
-
Scott - god
- Posts: 2384
- Joined: Wed Jan 17, 2007 4:43 am
- Location: in the queue for the ice-cream van
Re: LIDAR Base Maps
And I think those example maps show the problems of relying too much on the raw lidar contours. The 2016 map is much clearer and gives a good picture of the land shape. All the wiggles in the contours look as if they correspond to genuine land forms that would be recognisable on the ground. They have probably been slightly exaggerated so it is clear - this really is a spur or re-entrant etc. The new map has a lot of subtle undulations which makes the map more difficult to interpret and I doubt represent genuine landforms.
- pete.owens
- diehard
- Posts: 752
- Joined: Tue Jan 15, 2008 12:25 am
Re: LIDAR Base Maps
buzz wrote:I’ve read how some mappers produce a base map by taking every fifth contour, which clearly isn’t going to produce an good picture of the ground – missing out detail shown only on intermediate contours, but I wouldn’t have thought that’s any worse than photogrammetric plots or OS base maps.....Do most mappers produce base maps by taking every fifth LIDAR contour? ...producing a base map by tracing over the 1m contours, moving contours and adding form lines to pick out the smaller features and pick out the breaks of slope more accurately.
FWIW, I did my first terrain map for a decade or so for the Beecroft event back in September, remapping the area from scratch using LiDAR information (and the old map, and other bits and pieces!) for my base map. I didn't do any redrawing or prep of a single base map: I took OCAD Sketch out with me on a tablet, with various LIDAR layers and others set as background, and worked over the top of them. In terms of contours I had both the 1m raw contours and smoothed 2.5m contours to refer to. They were very different to the old contours, but I used both to inform me on shapes etc as I saw them on the ground - they were also invaluable in accurately positioning features in otherwise blank forest! Where I did stick to the LIDAR data was cross-referencing one part of the forest to another, i.e. making sure that ground shape mapped in one block was at the right level in relation to ground shape mapped in another block (which was the biggest weakness in the old contours). I tried to keep the resulting map's contours to within a metre of so of the LIDAR contours (not stuck to relgiously though), tending to use formline if the difference was more than a couple of metres and I needed to use the contour line itself to show shape.
I'm quite happy to be told if any of that was wrong by more experienced mappers - as I said it's a while since I did my last terrain map. Have to say that working with a tablet and with LIDAR was a revelation. The latter was particularly helpful in mapping vegetation, more so than contours (at least on the Beecroft map).
-
awk - god
- Posts: 3224
- Joined: Sun Mar 21, 2004 5:29 pm
- Location: Bradford
Re: LIDAR Base Maps
"However, the lidar contours had completely missed the adjacent dot knoll, which was already on the map and was patently obvious to anyone standing next to the hill. Presumably the interval between lidar readings was such that it had failed to take pick up the highest point of the knoll and whatever rise had been detected had been interpolated out when DSM was generated."
There could be another explanation. The source Lidar data is utterly huge - gigabytes or, more likely, petabytes of information from any flight. This is crunched extensively to produce the data that we then manipulate to produce contours.Part of this number crunching involves removing "noise" and I suspect that your dot knoll was regarded as "noise" by the algorithm and removed. This also explains why boulders, even fairly large ones, don't appear on Lidar contour maps.
I'm sure someone else with more technical nous can provide a more technical description of what I have skimmed over here...
There could be another explanation. The source Lidar data is utterly huge - gigabytes or, more likely, petabytes of information from any flight. This is crunched extensively to produce the data that we then manipulate to produce contours.Part of this number crunching involves removing "noise" and I suspect that your dot knoll was regarded as "noise" by the algorithm and removed. This also explains why boulders, even fairly large ones, don't appear on Lidar contour maps.
I'm sure someone else with more technical nous can provide a more technical description of what I have skimmed over here...
- Big Jon
- guru
- Posts: 1895
- Joined: Tue Nov 14, 2006 11:59 am
- Location: Dess
Re: LIDAR Base Maps
If you want dot knolls to show up on LIDAR, I'd be using the DTM, not the DSM.
As Jon says, possibly removed in the smoothing/interpolation algorithm if you're using DSM. it might also depend on if the data is 1m or 2m (this refers to the size of the grid square of the dataset as I understand it, but definitely *not* the contour interval).
Even so, no LIDAR dataset is foolproof, hence ground survey required too.
What LIDAR does do is (a) give you a better chance of getting the feature in the right place on the map instead of relying on pacing and (b) speed up your surveys.
As Jon says, possibly removed in the smoothing/interpolation algorithm if you're using DSM. it might also depend on if the data is 1m or 2m (this refers to the size of the grid square of the dataset as I understand it, but definitely *not* the contour interval).
Even so, no LIDAR dataset is foolproof, hence ground survey required too.
What LIDAR does do is (a) give you a better chance of getting the feature in the right place on the map instead of relying on pacing and (b) speed up your surveys.
- rf_fozzy
- light green
- Posts: 239
- Joined: Wed Feb 12, 2014 11:13 am
Re: LIDAR Base Maps
This isn't exactly a new problem though - photogrammetry also had all sorts of spurious bumps and wiggles. My go-to example for this is Dave Peel's map of Pwll Du from JK2014, which was a model of clarity compared to the apparently raw-PG predecessor.
I'm think much of the detail from lidar is "there, but should be omitted".
Working with 1m contours does also bring home how a contour on a good O-map needs to go uphill and downhill to capture the features (especially, given the "crag should be on a contour" requirement!)
I'm think much of the detail from lidar is "there, but should be omitted".
Working with 1m contours does also bring home how a contour on a good O-map needs to go uphill and downhill to capture the features (especially, given the "crag should be on a contour" requirement!)
WOC2024 Edinburgh
Test races at SprintScotland (Alloa/Falkirk) and Euromeeting (near Stirling).
Test races at SprintScotland (Alloa/Falkirk) and Euromeeting (near Stirling).
-
graeme - god
- Posts: 4727
- Joined: Thu Feb 12, 2004 6:04 pm
- Location: struggling with an pɹɐɔ ʇıɯǝ
Re: LIDAR Base Maps
Never use DSM for contours - this is a model from the first return pulses, so any vegetation (bush, scrub, tree etc) will affect it. Always use DTM for contours - but beware if they appear to form hexagonal lines - a sign that vegetation is so thick that the returns are giving garbage info.
DSM itself is fairly meaningless but subtract DTM from DSM and there is an image of vegetation that can be coloured to show height. Very useful for scattered trees or clearings in blocks of forest.
DSM itself is fairly meaningless but subtract DTM from DSM and there is an image of vegetation that can be coloured to show height. Very useful for scattered trees or clearings in blocks of forest.
- Big Jon
- guru
- Posts: 1895
- Joined: Tue Nov 14, 2006 11:59 am
- Location: Dess
Re: LIDAR Base Maps
You can process DTM LIDAR info into a "hillshade" picture and use that as a template. The hillshade can show knolls, depressions, ditches and earth walls and fences which don't appear in the contour processing. Trace the suspected features from the hillshade onto the base map and make sure you actually check it on a foot survey.
- Davy
- white
- Posts: 78
- Joined: Mon Jan 11, 2021 9:46 pm
Re: LIDAR Base Maps
As Jon said, automatically converting LiDAR contours is very bad practice, on both forest and urban maps.
As a relative beginner (~1 year experience) my method at the moment is to first draw contours at home, roughly tracing every 5th one but including obvious (big enough) features from the ones above and below, usually only if they show up on at least 2 or 3 LiDAR contours. Then I edit in the field based on what it looks like.
I'd be interested to hear whether people (experienced mappers) also do this or draw the contours from scratch while out in the field (and if so whether on paper or electronic)...
I recently remapped Hardcastle Crags and the contours on the old map (Stirling Surveys 1990) were like waves on the sea with amplitudes of 10+ vertical metres, in places where in reality the slope was pretty much flat! p.s. come to the event on 4th Feb
As a relative beginner (~1 year experience) my method at the moment is to first draw contours at home, roughly tracing every 5th one but including obvious (big enough) features from the ones above and below, usually only if they show up on at least 2 or 3 LiDAR contours. Then I edit in the field based on what it looks like.
I'd be interested to hear whether people (experienced mappers) also do this or draw the contours from scratch while out in the field (and if so whether on paper or electronic)...
I recently remapped Hardcastle Crags and the contours on the old map (Stirling Surveys 1990) were like waves on the sea with amplitudes of 10+ vertical metres, in places where in reality the slope was pretty much flat! p.s. come to the event on 4th Feb
- Alasdair Pedley
- string
- Posts: 12
- Joined: Sat Jun 29, 2019 8:24 pm
Re: LIDAR Base Maps
In OCAD I colour the 1m Lidar contours - thick purple lines for index contours, thin purple lines for 5m contours and v thin grey lines for 1m contours. Then using mapping film over the laminated basemap (to make waterproof) I draw the contours that I want to appear on the final map. Basically everything that is on the final map is on my field survey sheet. This is then scanned on a flat-bed scanner and registered onto the base map in OCAD.
Works well and any areas I have missed contours etc are easily spotted and can either be : A - revisited on a later survey trip, or B - added from basemap - if v thick area or v v steep.
I usually add a DSM-DTM model as well - shows vegetation. Plus occasionally I have produced a "local height difference" map - can't remember what OCAD calls it! Useful in picking up lines of paths and, often walls, fences etc.
All basemap's generated in OCAD11 - a bit clunky but does the job... eventually - to be fair its usually operator error if I have problems
Works well and any areas I have missed contours etc are easily spotted and can either be : A - revisited on a later survey trip, or B - added from basemap - if v thick area or v v steep.
I usually add a DSM-DTM model as well - shows vegetation. Plus occasionally I have produced a "local height difference" map - can't remember what OCAD calls it! Useful in picking up lines of paths and, often walls, fences etc.
All basemap's generated in OCAD11 - a bit clunky but does the job... eventually - to be fair its usually operator error if I have problems
- Big Jon
- guru
- Posts: 1895
- Joined: Tue Nov 14, 2006 11:59 am
- Location: Dess
Re: LIDAR Base Maps
w.r.t. removing the excess wiggles in Lidar contours, Open Orienteering Mapper has a "Simplify" function which seems to work quite well. I haven't found an equivalent in OCAD. Of course, the features still need to be checked on the ground.
curro ergo sum
-
King Penguin - addict
- Posts: 1453
- Joined: Wed Sep 21, 2005 6:56 pm
- Location: Kendal
Re: LIDAR Base Maps
King Penguin wrote: I haven't found an equivalent in OCAD.
Use the function that creates smoothed contours in the LIDAR import tool.
Can't remember exactly what it's called right now.
I usually will use these as my base contours, then adjust for features that show up on at least 2-3 of the 1m contours. Then add form lines as appropriate. (For ISSprOM - for ISOM, I'd do 5m contours with 2.5m form lines obviously)
I then check on the ground and see if the form lines can instead be incorporated into the main contours or are not required.
In general I also now only print the drawn map (as adapted from initial basemaps) onto paper and use a weather writer to draw directly onto the sheet. Then scan this to adapt my map.
I also try, at least for ISSprOM maps to do as much in a single survey as possible. I only return if I really need to check something or the area is particularly complicated.
For ISOM, obviously need more survey time, but then less cartography time.
- rf_fozzy
- light green
- Posts: 239
- Joined: Wed Feb 12, 2014 11:13 am
18 posts
• Page 1 of 2 • 1, 2
Who is online
Users browsing this forum: No registered users and 13 guests