About the Z Threshold
The Z threshold parameter represents the maximum orthogonal distance
to the
plane under consideration
required for a point to be classified. If the plane is a good fit,
based on the
plane fit
parameter, new points are added to the plane during the
surface
growing if the points are less than the Z threshold value.
The threshold value depends on the accuracy of the input point cloud
data. In addition, the object of interest under consideration and
the quality of the point cloud data would affect the value to use
for this parameter. For example when using the
planar
filter to classify buildings, certain roof types might require
a higher value for this parameter such as roofs with Spanish tiles
where there might be significant surface roughness on the roof. If
there is poor alignment amongst various LIDAR tiles or flight lines
a higher value might be required. Finally, the value used for this
parameter would typically not be much larger and not greater than
two times the value input for
plane
fit. The Z threshold value is linear and the default value for
this parameter is appropriate for most cases.
Effects of Increasing or Decreasing the Z Threshold
- Decreasing Value: Decreasing the Z threshold value will result
in fewer points being classified. Lower values will result in
fewer classifications because the points under consideration are
required to be a 'tighter' fit to the given plane resulting in
fewer points being classified.
- Increasing Value: Increasing the Z threshold value will result
in more points being classified. Higher values will result in
more classifications is because the points under consideration
are allowed a 'looser' fit to the given plane resulting in more
points being classified.
