Grid_Terrain In absence of class 2

[nidhi23ug]

First just to mention, I have used liDR extensively all through my doctoral thesis and I can not thank you enough for such great and easy tools.

There has always been one issue that I wanted to raise is the issue of Grid_Terrain when there is no class 2. So one data provider did not follow LAS specifications and classified ground points in class 1 and 2 as unclassified points, and in some tiles, there are no unclassified points. So obviously in some tiles, there is no 2.

However, when I use class 1L for grid_terrain, for some tiles where 2L is missing I keep on getting "No ground points found. Impossible to compute a DTM." I have checked the source code and I think this is forced in even if class chosen for interpolation is different.

I know how to surpass it, I just kept Null class as 2 and I can still run it. But I wish I do not have to do it.

[nidhi23ug]

lidDat = readLAS(Path)
unique(lidDat@data$Classification)
#> [1] 6 1 8
dtm= lidR::grid_terrain(lidDat,res = 0.25,algorithm=knnidw(k=10L),use_class = c(1L))
#> Error: No ground points found. Impossible to compute a DTM.

[Jean-Romain]

Try to use -change_classification_from_to 1 2 in filter. This will reclassify at read time

readLAS(path, filter = "-change_classification_from_to 1 2")

[Jean-Romain]

Also

grid_terrain(lidDat,res = 0.25,algorithm=knnidw(k=10L),use_class = c(1L, 2L))

Should do the job no?

[nidhi23ug]

use_class = c(1L, 2L))

No, this one does not do, as there is no 2 class in some. Yes, reclassification does help. I already tried doing the reclassification and it worked fine.

[Jean-Romain]

If there is no class 2 then only class 1 will be used. I don't understand why use_class = 1:2 cannot do the job.

[nidhi23ug]

If there is no class 2 then only class 1 will be used. I don't understand why use_class = 1:2 cannot do the job.

Yes, this is precisely why I raised the issue. I could always reclassify but I didn't understand why it does not work.

`> lidDat = readLAS(Path)

unique(lidDat@data$Classification)
[1] 6 1 8
dtm= lidR::grid_terrain(lidDat,res = 0.25,algorithm=knnidw(k=10L),use_class = c(1L, 2L))
Error: No ground points found. Impossible to compute a DTM.`

[Jean-Romain]

Ho ok I got it. Will see that tomorrow. It looks like a bug. Are you able to share a small file that I'll be able to use to reproduce.

[nidhi23ug]

Ho ok I got it. Will see that tomorrow. It looks like a bug. Are you able to share a small file that I'll be able to use to reproduce.

Yes sure, can I directly send it to you without keeping it public?

[Jean-Romain]

As you wish

[Jean-Romain]

Fixed. You can install from github or wait few more day I will release v3.0.1 soon.

The problem was that when I introduced use_class options I didn't remove a test consting in counting the point classified 2 to fail early in absence of point. This invalidated your specific use case.

Thank you for reporting

[nidhi23ug]

Fixed. You can install from github or wait few more day I will release v3.0.1 soon.

The problem was that when I introduced use_class options I didn't remove a test consting in counting the point classified 2 to fail early in absence of point. This invalidated your specific use case.

Thank you for reporting

Ah thank you so much for your quick action. Will do. Cheers