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

Discussion: KP batch processing

in: Orienteering; General

Oct 9, 2020 1:22 AM # 
robplow:
I'm doing some stuff in KP but when processing tiles individually I get a white gap around each tile:


would batch processing solve that?

So then the questions is how to do batch processing - I can't quite figure out what to do from the instructions in the read me file.
Advertisement  
Oct 9, 2020 2:08 AM # 
Spike:
Tutorial with KP batch processing linked from here:

http://www.attackpoint.org/viewlog.jsp/user_29/per...
Oct 9, 2020 2:25 AM # 
robplow:
Thank you!
Oct 9, 2020 2:30 AM # 
Canadian:
And yes I believe the batch processing solves that.
Oct 9, 2020 6:45 AM # 
Jagge:
robplow, consider also setting savetempfiles=1 in ini file (at the end of ini file). That makes it not store only png images but also those vector files like dxf contours. And settinig something for "basemapinterval" will make it produce any contour interval dxf you want.

If you process very large area with lots of tiles you can make buffering faster by indexing laz fiels with lastools lasindex before starting the batch.
Oct 9, 2020 7:11 AM # 
robplow:
Thanks - I have four tiles processing now as my first attempt at this.

At this point i am just trying to get an idea of whether some areas I have found lidar for have potential as O maps. Not really trying to produce training map or base map so the png output is quite sufficient.

next step will be trying to adjust the green settings.

BTW: the lidar I have was flown 2019 - over 20 points per m2 on some tiles (according to OCAD DEM wizard)
Oct 9, 2020 7:37 AM # 
robplow:
Those four tiles are done - you are correct Jeff - batch processing did eliminate the white tile borders!
Oct 9, 2020 11:46 AM # 
robplow:
Here you are Jagge - this a comparison of an area I mapped in 2017 using 2012 or 14 lidar with KP output using 2019 lidar. Actually Jeff (Canadian) did the SE part in 2015. I haven't adjusted the vegetation settings at all. Pretty good!

Oct 9, 2020 11:49 AM # 
tRicky:
That doesn't look at all like Kings Park.
Oct 9, 2020 12:49 PM # 
robplow:
A better comparison - with the black and blue features overlaid on the KP output

Oct 9, 2020 12:58 PM # 
Canadian:
@robplow, neat to see that. Theres some cool stuff in area!
Oct 9, 2020 3:58 PM # 
Jagge:
I believe that's good enough to study o map potential even if light green isn't as light as it should. If there is enough white blobs in green you can suspect may be light green. That's somewhere in Canada? Edit: right, I spotted the location.
Oct 9, 2020 4:00 PM # 
Canadian:
Yes, it's an area called Debert in Nova Scotia: https://www.google.ca/maps/place/45%C2%B024'37.6%2...
Oct 10, 2020 2:24 AM # 
robplow:
And just to compare - here is the same area KP output for the older lidar (2013). I tried the building generator as well.

Oct 10, 2020 6:12 AM # 
GuyO:
Was this area used on the last day of the 2015 COC/O-Fest?
Oct 10, 2020 10:52 AM # 
Canadian:
Guy, yes it was.

Rob, were the KP settings the same for the runs of the two different lidar data sets? Its interesting to me that the older data comes out with a generally lighter set of green shades.

If you didnt change the KP settings would that be because the newer data has a higher point density?
Oct 10, 2020 12:19 PM # 
Jagge:
Scanning date makes difference, also flight elevation, scanner hardware/software etc can make difference in sensitivity of picking internediate returns between first and last.
Oct 10, 2020 9:11 PM # 
robplow:
Jagge, can you suggest some tweaks to the vegetation settings to make the green in the image from the 2019 lidar match the map better?

Jeff, same settings in both, except for adding buildings in the 2013 lidar output

the 2013 was 4 points per m2, the 2019 stuff 21.
Oct 13, 2020 12:53 PM # 
Jagge:
It depends what is seen as being wrong. It it is just color tone, that can be tweaked with photo editor. If it is not generalized enough you can add some more median filtering (and loose some deatails). If undergrowth stripes are too annouying you can raise those figures triggering them, at the beginnig of ini file I think. If there is some spesific errors - some type of vegetation being mapped wrong, then one needs to look at lidar footprint, for example point cloud profile images, and figure is there returns from that type of vegetation and how to count them in with more (or less) weight. So, not necessarily so straight that forward to give easy suggestions because unfortunately(?) it's already pretty good for being a KP vegetation output.

Please login to add a message.