User Tools

Site Tools


mapping:automatedmapping

Automated Mapping

Systems that bypass fieldwork. Kartapullautin is the pioneering software here but others will presumably appear.

MT: mcroxford Posted: 6 June 2012, 5:31 PM

Is anyone else following the Facebook thread of Routegadget at the moment? We have sourced a lot of Lidar data recently from our local Council and following his process with interest.

MT: mcroxford Posted: 9 June 2012, 6:20 PM

I think so. But also he is applying some other programmes which when there is a certain closeness to the contours then it is interpreting them as cliffs but more interesting when they capture Lidar there is top of canopy info and also bottom of canopy info. He seems to be using the difference between the two to map vegetation. Where the valley are essentially the same then no vegetation but when there is a differential then he's mapping the vegetation. Also there seems to be some sought of capture of density of the vegetation which he is using to determine runability.

MT: Michael Posted: 9 June 2012, 11:38 PM

Ah, the holy grail. Push a button and the map is made:-))

Actually, I think that some of Stewart Hyslop's basemaps could be run on without any fieldwork, where the vegetation is consistent eg white forest.

MT: mcroxford Posted: 14 June 2012, 8:38 PM

http://okansas.blogspot.co.nz/2012/06/more-fun-with-lidar-and-vegetation.html and http://iso-henna.com/doma/show_map.php?user=Jagge&map=104 to read more on the holy grail approach to map generation… Perhaps more alchemy?

MT: richardh Posted: 20 June 2012, 8:02 PM

Latest from routegadget. Will be great if he releases his processes for generating a base map. http://www.facebook.com/pages/RouteGadget/177518995597572

MT: mcroxford Posted: 22 June 2012, 6:49 PM

MT: mcroxford Posted: 5 November 2012, 5:27 PM

First event (I think) on a Pullautin map. What's that? It is a computer generated map using Lidar data and some other sources in Finland. To be used for Ski-O I think from the poor translation from Google. Check out the size of the map! I'm assuming a 1:10,000?

http://www.vaajakoskentera.com/sadismi2012/

MT: rossmaxmo Posted: 5 November 2012, 7:01 PM

Cool eh?? It's actually being for an ultralong distance race called 'Sadismi' - I'm sure you can guess what it means. The map is generated by software created by the dude from Routegadget from an xyz file, at the moment it can give pretty good enough maps for training, with pretty accurate contours, cliffs, rocks and even knolls and vegetation. There's a video by Jan Kocbach running on one of the generated maps here: http://youtu.be/TFFW_QfkosU

You can get the software here: http://routegadget.net/karttapullautin/

MT: rossmaxmo Posted: 5 November 2012, 7:02 PM

*being made for an ultra-long (foot orienteering)

MT: Martin Posted: 5 November 2012, 9:49 PM

I've been playing around with pullautin and it's pretty powerful.

In Auckland the rural tiles don't have a very high density of lidar points. I also didn't have the above ground points for this sample of the Kohekohe map, so some of the finer contour detail hasn't come through. http://cmoc.co.nz/lidar/Kohekohe_Mapped.jpg http://cmoc.co.nz/lidar/Kohekohe_Lidar.png

Another map in progress… http://cmoc.co.nz/lidar/quarry_lidar.png

MT: mcroxford Posted: 6 November 2012, 6:00 AM

Behold. The Holy Grail. Sorry Michael, couldn't help myself. :-)

MT: Michael Posted: 6 November 2012, 1:31 PM

Fascinating! I suppose the inference of non-contour features might have been calibrated on Scandinavian (?) terrain, how well do you think it's detecting NZ features? Nice to see the absence of U depressions given how few are really 1m deep, but I wonder if the knoll-detector isn't a tad over-sensitive. Would be great to assess these in the field (no doubt you have…)

MT: Martin Posted: 6 November 2012, 1:40 PM

The dot knoll and small depressions haven't come out particularly well in the Kohekohe example. I think that is because of the low density of the points. The auckland data is 1 point per 25 square meters for rural data and 1 point per 2 square meters for urban data. In comparison the data that the is being used for development is up to 6 points per meter square.

The patch of detail in the trees also didn't show up much. This was probably because I didn't have the above ground points.

So essentially the output is only as good as the input data.

MT: Martin Posted: 6 November 2012, 1:42 PM

Looking at the Quarry sample, it's picking up areas of vegetation but some very nice white forest is showing up green. At a guess because it has a denser canopy?! The levels of green can be tweaked so that's the next step

MT: Michael Posted: 18 December 2012, 4:35 PM

And another link from Ken about vegetation mapping using LIDAR data. Also by Thomas Gloor, as above. LIDAR data reflected from the ground and the vegetation top are subtracted to give a vegetation height map. Not as automated as Pullautin, the vege is printed as colours for various heights, along with contours etc to provide a base for a person doing fieldwork. Done any more experimentation Martin? Anyone else? http://ocad.com.au/wp-content/uploads/2012/12/Vegetation-Height-Maps-OCAD11-201211.pdf

MT: Martin Posted: 18 December 2012, 5:44 PM

In the process of getting some more data… Watch this space Michael

MT: Michael Posted: 4 June 2013, 11:12 AM

Martin used Pullautin to generate the basemap from raw LIDAR data for yesterday's map “Bombay Quarry”. A fairly straightforward spur-gully area but nevertheless this is a significant achievement. (Jason I think you were generating contours from the raw data at Teetotal but perhaps not inferring vegetation differences from the treetop reflections?)

MT: Linley Posted: 23 August 2014, 1:55 PM

This interesting snippet came through from up north. Might be of interest to NZ mappers if you haven't seen it already. With the cold winds blowing us here in Christchurch, indoor mapping has a strong appeal! o-zeugs is keen for feedback. http://o-zeugs.blogspot.no/2014/08/couch-mapper-from-las-to-map-without.html?m=1

MT: Martin Posted: 26 August 2014, 8:46 AM

I've been playing around with karttapullautin for a while and produced a few maps using it.

The latest mission is some fairly extensive coverage of woodhill and riverhead.

There's many parameters to test which control dot knolls, cliffs, smoothness of contours and vegetation. Green is used to represent different height of vegetation, rather than strictly being run-ability.

Here's a snippet from woodhill (Restall Road/Inland Road): http://cmoc.co.nz/wp/wp-content/uploads/2014/08/Woodhill_sample.pdf

Kenny 17 Nov 16 0850

The coming summer edition of The Australian Orienteer has an article on Karttapullautin experience by visiting coach and mapper Stefano Raus assisted by OA's Adrian Uppill. More on that and two other mapping technology articles at ocad.com.au/blog

Michael 2 Dec 16 0037

This is amazing: Finnish mappers have used Karttapullautin, running on many computers in parallel, to produce an orienteering map of the entire country! Read about it at http://www.mapant.se/fi/about.php

mapping/automatedmapping.txt · Last modified: 12:50pm Tue 25 October 2022 by 127.0.0.1