Register | Login
Attackpoint - performance and training tools for orienteering athletes

Discussion: Democratized Maps

in: Orienteering; General

Nov 2, 2019 2:18 PM # 
Gswede:
Here's a new project. Democratized Maps. Autogenerated maps in the US that are available for download and use by anyone. Making these maps is pretty useless if it stops with me.

Right now there are only three maps in Northeast Pennsylvania. But I'll continue to expand it as I develop what I feel our usable maps. Let's enjoy our sport!

Democratized maps
Advertisement  
Nov 2, 2019 3:02 PM # 
Terje Mathisen:
Jarkko + Mats &co created mapant.fi several years ago, Sweden is also ready but waits on some form of government permit. Here in Norway we have done 90% of the ground work and are ready to start making mapant.no tiles.

1:10K, ISOM2017, will cover all of Norway in a couple of years.
Nov 2, 2019 3:03 PM # 
Gswede:
Yep! This is an attempt to start emulating that stuff. But we have to adjust it for the US because of our private property culture.

Edit: now I actually have time to write. Many thanks are due to the people who created Mapant and the projects you mentioned. They provided the inspiration for me to start doing this. We'll see where it goes!
Nov 2, 2019 4:11 PM # 
BrianJohnston:
Nice work.
Nov 3, 2019 12:40 AM # 
gordhun:
That is neat.
So let's say someone takes one of those maps and goes to hold an orienteering event. She goes out to check the terrain and finds a couple of corrections that have to be made - say those purple lines are actually contours but really represent depressions or there is a fence or a marsh missing and would like to add them.
How does she go about doing that?
Nov 3, 2019 1:39 AM # 
tRicky:
She is a girl so she is Wrong.
Nov 3, 2019 1:53 AM # 
bmay:
I presume that was meant to be funny. It wasn't.
Nov 3, 2019 2:20 AM # 
Gswede:
Good question, Gord!

If someone were to be that interested in updating the map and felt they had the skills to do it, then they should just contact me for the OCAD file.

But really, these maps shouldn't be made into competition level maps, except for perhaps local events and events outside of national competition, like a goat event just for fun. For official competitions, we need to continue valuing the work done by experienced mappers.

Autogenerated maps have a place in the modern world of orienteering, but they can't replace a proper fieldchecked map.
Nov 3, 2019 3:40 PM # 
Terje Mathisen:
I have mapped 200+ sq km in Hvaler, but only 12-13 sq km (the Asmaløy island) has been properly surveyed, as a (very large!) sprint map, the rest is all autogenerated plus 6-7 years of hiking around combined with path system survey.

https://tmsw.no/asmaloy/
https://tmsw.no/hvaler/

I release both the original ocad files and a reasonable resolution image version of these maps.
Nov 4, 2019 2:56 AM # 
Gswede:
Man, those look fun.
Nov 4, 2019 2:52 PM # 
gordhun:
GSwede: Back to those PA maps. I was sitting around with not much to do so I thought I'd try to see what output could come with OCAD.
It started out well. The Mocanaqua detail on Open Street Map is very good.
The National Map Viewer Download has tiles for that area but as I've learned to do I check the tile meta data to see what projection system has been used. In America it can be one of three or more systems.
When I find UTM that is like getting a Get Out of Jail Free card in Monopoly.
State Plane? Okay OCAD can handle that automatically, too. But these PA files are Lambert Conformal Conic which OCAD can also handle if the right info was inputted and as I recall the last time I tried for a map in Virginia a key beit of information was missing.
So the question: How do you handle this projection so you get a geo-referenced map?
Nov 4, 2019 2:57 PM # 
JLaughlin:
State Plane is a Lambert Conformal Conic.
Nov 4, 2019 3:00 PM # 
edwarddes:
Reprojecting data is a basic part of GIS work.
If you like UTM, then just reproject everything to UTM. Want state plane? Reproject to that.
Nov 4, 2019 3:06 PM # 
Gswede:
Yeah, the trails from OSM on Mocanaqua are really in accurate based on my trips to the terrain. They're probably decent for hiking, but definitely not orienteering.

And I usually just reproject the data to sp83 PA_N with meters as the target units. That seems to do the trick.
Nov 4, 2019 3:25 PM # 
gordhun:
Trails? You want accurate trails?
Walk in the shoes of those that have gone before.
Here is the image from Strava Heatmap of the western part of that Mocanaqua map. I find the Strava heat maps far more reliable than any solo effort of a public agency or someone alone inputting to Open Street Map.
Nov 4, 2019 3:27 PM # 
Gswede:
Heat maps is pretty good, but I can actually see all the trails, including the long unused trails with hillshades and slope images.

It is fun to see some of our orienteering tracks on that image.
Nov 4, 2019 4:25 PM # 
gordhun:
Yes, I live in fear of some park officials seeing a heatmap of their park with the tracks frozen in time and thinking that there is way too much off trail activity in their park. Doesn't matter that the thin pink line is one snapshot of one day perhaps several years ago.
Nov 4, 2019 4:49 PM # 
hughmac⁴:
Nice work, Gswede!!
Nov 5, 2019 7:54 AM # 
Terje Mathisen:
@Gswede: I always use the slope image to verify all trails and paths. The major ones show up perfectly while even the very thin trails can be verified as to exact location when you already know that there is a path there.
Nov 5, 2019 12:01 PM # 
Canadian:
+1 Terje. Strava gives a good start but doesn't compare to Lidar imagery when it comes to precise placement.
Nov 5, 2019 2:02 PM # 
gordhun:
I agree with Canadian but I'm just not as confident as they are at seeing all the trails with LiDAR. To me Strava gives a better place to go looking and field checking than any of the other on-line sources like Google Earth or Open Street Map.
But to each his own.
Nov 5, 2019 2:21 PM # 
Terje Mathisen:
@gordhun: We're in violent agreement here!
When you have significant strava activity, the heat map shows very clearly where there must be paths: You then use the lidar slope image to adjust the strava trace for exact placement of those paths.

We still have many, many areas, even close to Oslo, where most paths except the largest ones, have effectively zero strava tracks, or where all the off-main-paths traces are from orienteers.

There is no way to determine of this was an off-path route choice taken by 3-4 runners, or an unmapped path, so this has to be verified while surveying, but you need to do that anyway to determine the proper classification, right?
Nov 5, 2019 2:35 PM # 
Spike:
I've had good experience using "sky view factor" to pick out trails from lidar using the Relief Visualization Tooldbox:

https://iaps.zrc-sazu.si/en/rvt#v
Nov 5, 2019 5:02 PM # 
Terje Mathisen:
All the various experiments on picking up depressions/hillocks and linear traces perform the same, within a factor of two or so: The important issue is to convert the very noisy near-ground/ground points into something which is visually much less noisy. Our eyes are really good at spotting patterns. :-)

Please login to add a message.