Autocorrelation for Raster Data

(currently only .BIL-Files)

Option 1: Equal distributed probe points or within a hull (.SHP-File)

  1. Generate a grid of probe points NxN
  2. Probe value (e.g. elevation) Z(x,y) in each DEM file (An optional SHP-File with an area feature can be used to specify the examination area)
  3. Extract grid values (e.g. elevation Z), in distances s_i = (0 … # of rangerings) and direction d_i = (North, South, East, West … ) and each Point ID: #Point
  4. dZ(#Point,d_i,s_i) = Z(#Point,d_i,s_i) - Z(0)
  5. Build Meanvalues for distance s over all points and directions
  6. dZ_Mean(s_i) = SUM(dZ(#Point,d,s_i)) / (4*NoOfPoints)
  7. Calculate the Autocorrelation for the
  8. dZ_Mean(s) values

Option 2: Along an axis, and orthogonal to the axis at every axis line point.

Works like Option 1 except:

  • Directions di are along and orthogonal to the given axis (.SHP-FILE with line)
  • probe points are taken along axis

Output:

	Mean Values:
	distance to point; r_autocorrelation; # of valid points

Comments are closed.