It is possible for instance predictions to contain fewer points than are present in the original points and semantic segmentation data. After outer joining, this can result NaNs in the PredInstance column of the merged dataframe.
In lieu of a potential fix of the root cause of this, this update simply assigns 0 to all missing values in the PredInstance column after merging (prior to saving to .las format).
It is possible for instance predictions to contain fewer points than are present in the original points and semantic segmentation data. After outer joining, this can result
NaNs in thePredInstancecolumn of the merged dataframe.In lieu of a potential fix of the root cause of this, this update simply assigns
0to all missing values in thePredInstancecolumn after merging (prior to saving to.lasformat).