Outlier

Detection of Single Pixel Outliers in DEMs using a static or a dynamic thresholding method. Fills the outliers with mowing window data from the surrounding or writes a NODATA value (-4444). Reference: Hengl, T., Reuter, H.I. (eds) 2008. Geomorphometry: Concepts, Software, Applications. Developments in Soil Science, vol. 33, Elsevier, 772 pp. in chapter "Preprocessing" http://dx.doi.org/10.1016/S0166-2481(08)00004-4

INPUT

Input Data - Elevation DEM (Raster Dataset) - a grid representing a continuous surface

COMMAND

Command Call

Type of Threshold:

  • Static Threshold: The Standard Deviation Threshold is set to a value and is constant over the whole dataset.
  • Dynamic Threshold: The Standard Deviation Threshold is calculated based on the Standard Deviation in the Window times the Value given in Standard Deviation for Outlier Values.

Standard Deviation for Detection:

  • Case Dynamic Thresholding: In smooth DEMs a value of 2, in rougher/noiser DEMs higher values are needed.
  • Case Static Thresholding: Specify the Standard Deviation Value which is supposed to be an outlier (e.g. 20 or 40)

Extent Search Window:

  • The window for the Standard Deviation Generation. Large window Sizes increase computation times.

Outlier Value:

  • Instead of filling the dataset by a moving window approach a nodata value (-4444) is written to the output grid.

RESULTS

Output DEM (Raster Dataset) - a grid representing a continuous surface

Dynamic 2 SD outliers with average replacement

Dynamic 2 SD outliers with extra ERROR value written

Static threshold of 5m with average replacement

Static threshold of 5m with extra ERROR value written