<div><div dir="auto">The options are very heavily borrowed from the PCL library, where this was a squared value. Also, keep in mind that the default values may be better suited for terrestrial/robotics applications (as opposed to aerial), and will very likely need some tweaking. </div></div><div dir="auto"><br></div><div dir="auto">The reported transformation should also be reported in row major order, like the transformation filter. If you find it is not, I’d consider that a bug, so feel free to file a ticket. </div><div dir="auto"><br></div><div dir="auto">PRs are always welcome, especially for documentation!</div><div dir="auto"><br></div><div dir="auto">Brad</div><div><br><div class="gmail_quote"><div dir="ltr" class="gmail_attr">On Tue, Feb 18, 2020 at 19:11 Joachim Meyer <<a href="mailto:j.meyer@utah.edu">j.meyer@utah.edu</a>> wrote:<br></div><blockquote class="gmail_quote" style="margin:0 0 0 .8ex;border-left:1px #ccc solid;padding-left:1ex">Hi PDAL-devs,<br>
<br>
I have a few questions using `filters.icp` (<a href="https://pdal.io/stages/filters.icp.html" rel="noreferrer" target="_blank">https://pdal.io/stages/filters.icp.html</a>).<br>
First off, I am very new to aligning two point clouds using this filter and have a limited understanding of the underlying algorithm or terminology.<br>
<br>
The options give a default value of 9e-8 for the translation threshold.<br>
Looking at the source code, I read that if I wanted to have a different value for that, I should give the square root for a different threshold: <a href="https://github.com/PDAL/PDAL/blob/a8b0bfbefb7f8ffc467092f11cedc155f862fa6f/filters/IterativeClosestPoint.cpp#L68" rel="noreferrer" target="_blank">https://github.com/PDAL/PDAL/blob/a8b0bfbefb7f8ffc467092f11cedc155f862fa6f/filters/IterativeClosestPoint.cpp#L68</a><br>
<br>
Is that correct?<br>
<br>
My use case aligns two larger point clouds (100+ million) to each other and, to improve performance, I reduced the size of those through downsampling in point density, as well as filtering to stable reference ground between the two. I save the output of that alignment and now planning to use the same matrix with `filters.transformation`. There, it says that it should be given in row-major order, and I assume that is what the output of `filters.icp` with the —metadata is giving me?<br>
<br>
Let me know if I can contribute your answers back to the docs through a PR.<br>
<br>
<br>
Thank you,<br>
Joe<br>
<br>
_______________________________________________<br>
pdal mailing list<br>
<a href="mailto:pdal@lists.osgeo.org" target="_blank">pdal@lists.osgeo.org</a><br>
<a href="https://lists.osgeo.org/mailman/listinfo/pdal" rel="noreferrer" target="_blank">https://lists.osgeo.org/mailman/listinfo/pdal</a></blockquote></div></div>