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

Discussion: Karttapullauitin Question

in: Jagge

Jan 7, 2016 2:01 PM # 
Moi Jagge,

I was in contact with GSwede about how your system works and he redirected me to you. I was wondering on what basis the terrain types are determined and how the overall system works? As in how does the program know whether it is runnable or fight or open? Im looking at doing a lit review on this type of study and would be greatful if you could explain or send me some information about the workings of it so i can reference your work. if you get a chance drop me an email:

Jan 7, 2016 2:43 PM # 
Well, it all is based lidar point clouds. Other apps must have classfied some of the points as ground first. Those ground points are used for contors and cliffs. And basically rest is vegetation. It sort of counts how many points given area has in total and how big % of those are 0.3 ... 3 m from ground. Higher the % gets, more green it is. And if tehre is no or almost no points above ground it is yellow.

Of course it is not entirely that simple, for various reasons. That is why there is plenty of parameters, like meter from ground values need to be set differently for different vegetation types and datasets. And there is some problems and issues like dead zones and workaround parameters for these and that makes it not that straight forwards and difficult to explain (and also tune right).

Google "lidar could forest" images and you should get idea how those points clouds may look like.
Jan 7, 2016 2:55 PM # 
For my MSc thesis i was looking at designing a map generation software (i guess similar to this) which would use an image input and a LiDAR basemap in order to generate an output of an orienteering map designed to IOF standards and then build on it and compare it to field drawn maps etc. Would you have any advice?
Jan 8, 2016 7:21 AM # 
I'd first compare existing O maps and aerial images and try figuring out what can and what can not be detected from the imagery. and what kind of imagery is needed to detect something usefull. And build on that. And not worrying if something essential can't be detected, it is enough to somewhat reliable detect something, it may complement nicely lidar mapping or make manual mapping faster.

This discussion thread is closed.