ONF auto DBH detection -- some questions
I’ve been working with the ONF DBH tutorial: http://rdinnovation.onf.fr/projects/computree/wiki/En_dbhAuto_.pdf
So far, the stem maps are looking pretty good but I had a few additional questions:
1) Is there an easy way to “recover” the original x,y,z position since I was using the “Center Point Clouds” function in case some PCL issues arose? The output csv is in the centered coordinates, but that makes it impossible to compare against a “true” stem map.
2) Is there a way to get the predicted tree height from this analysis, or do I need to do another analysis afterwards to get there. Can I take the seeds from this step and move it to the SimpleTree step? I’m finding at least for my applications that the seed detection from the ONF approach (with the filtering) seems to be doing better than Simpletree’s.
3) Can I export the cylinders somehow to CloudCompare for better visualization?
My solution for 1) would be to center the point cloud using some kind of python script or other software where you can note the translation applied. You could then apply the inverse to your stem map. Shameless plug, if you can’t do that, you can align the stem map produced by Computree with your reference stem map using my work: https://github.com/jefft255/TargetlessForestRegistration (more detail in the paper https://www.sciencedirect.com/science/article/abs/pii/S0924271618302892) but it would be really better if you used the translation.
2) You can probably move some step around and do that, although I couldn’t tell you exactly how.
3) No idea!