A demonstration of LP360’s Model Keypoints algorithm.
The Model Key Point Point Cloud Task can be summarized best if you consider all the points in your DEM and for each one, remove the point and look at the amount of deviation that occurs with the surface. If that deviation exceeds the the limit you set then the point is considered a Model Keypoint and flagged accordingly since the surface would deviate more than the allowed tolerance from the original surface if the point is not used. The second criteria, sample distance, has a much smaller impact on the final number of keypoints, but simply says if the distance on a flat surface exceeds that sample distance, then a point even though it doesn’t exceed the deviation limits must be considered a keypoint as well. You can think of this second criteria as the maximum allowable grid size. Think of a very large flat parking lot with no elevation variation. We may wish to define a minimum grid spacing even though the entire surface of the parking lot could be defined by four points. For example, we may wish to have a point every 100′ at a minimum.
MKPs will occur in greater density where there is elevation change and less dense in flat areas. If you have noisy data, using a small deviation value can cause more MKPs than desired.