To the list of courses || RASA2018 || To the theme || Estonian

Exercise 3992. Points 3, theme: Local Statistics (numerical variable)

Open exercise
The file Madsa_L.rst in the attached archive contains a greyscale ortophoto where pixel edge is 1 m.
  1. Calculate distance weighted autocorrelation of pixel values in a 10 m grid within round local kernels. Present the result as a figure.
  2. Where is the autocorrelation larger and where is it less?
  3. Explain, how different pixel values determine the differences in autocorrelation values in these sites.


The ortphoto image at the Madsa Sports Center is used in this excercise.
  • Extract files from the attached archive.
  • Open calculation of Local statistics.
  • Select Upload rst file, numerical layer, grid over input layer, Parameters from rdc.
  • Leave Void value empty or write 255. Actually, there are no void areas in the input file.
  • The Output grid interval (units as in coordinates) should be 10 as the coordinates are in meters and the output grid is asked to have 10 m cell interval.
  • The kernel internal radius is 0 and external radius 10. NB! In case of a larger radius the calculation time will overflow.
  • The statistic to calculate is reverse distance weighted Moran I.
  • NB! Select the rst and rdc files to upload as the last step because upload is valid only during single postback.
Log in to send your results and to see the expected answer and responses from other students.