[pdal] Digitized building footprints

[Stephen V. Mather]

Ha! Ok, my gut was right. Rasters it is.

And by massive, I just mean 80k images worth of photogrammetric point clouds over 300km2. I know I estimated how many points that is at one point, but I don't recall now. 😄

[http://sig.cmparks.net/cmp-ms-90x122.png] Stephen V. Mather
GIS Manager
(216) 635-3243 (Work)
clevelandmetroparks.com<http://www.clemetparks.com>



________________________________
From: Howard Butler <howard at hobu.co>
Sent: Monday, June 29, 2020 12:50 PM
To: Stephen V. Mather <svm at clevelandmetroparks.com>
Cc: pdal <pdal at lists.osgeo.org>
Subject: Re: [pdal] Digitized building footprints


WARNING: This email originated outside of Cleveland Metroparks. DO NOT CLICK links or attachments unless you recognize the sender and know the content is safe.


On Jun 29, 2020, at 10:18 AM, Stephen V. Mather <svm at clevelandmetroparks.com<mailto:svm at clevelandmetroparks.com>> wrote:

Hi All,

I am processing a dataset for which I have a digitized building footprints. It'd be even cooler if I had vegetation, but beggers != choosers.

In principle, I could use filters.overlay, but this is a massive dataset, so I am curious if it makes more sense to rasterize, tile things up, and use filters.colorization.

Define massive ;)

I think your instinct to rasterize the footprints is the right one here. filters.overlay is going to do point-in-polygon for every point/poly combo. Most are quickly thrown out, but you're going to be checking every one. A raster mask is going to be much quicker, and you can control the resolution with gdal_rasterize.

Happy to test and report back on what is fastest, but in case someone has a hunch, or an alternative recommended approach, I thought I would ask.

I'm curious in how different things might be, but I suspect the approaches will separate quite quickly for significantly sized polygon and point cloud sets.

Howard

-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.osgeo.org/pipermail/pdal/attachments/20200629/a11b1c55/attachment.html>