2

I am interested to normalize Landcover raster image having 6 classess using fuzzy membership function (sigmoidal monotonically increasing) in Weighed Linear Combination (WLC) in the backdrop of MCE approach.

The aim is to produce Landslilde susceptibility map. The weights will be derived using pairwise comparison of the attributes to get their relative importance over one another.

Normalization/standardization of input criterian maps is a prerequisite in WLC. I have already ranked my classes from 1-6. Let me know how could I assign inflection points to my landcover classes on 0-255 normalization scale to produce my fuzzy-landcover map?

  • so you already have a raster and you just want to scale its values to go from 0-255? what software are you using? python? – user1269942 Apr 18 '16 at 2:37
  • I am using IDRISI Selva 17.0. The aim is to standardize the Landcover raster having 6 classes to 0-255 scale but not sure how to assign inflection points to prioritize my classes. – Ben Apr 18 '16 at 5:57
2

Landcover classified image values (discrete classes) could be scaled from either 0-1 or 0-255 (continuous scale) by Ranking them in accordance to their relative importance for particular purpose. In may case landslide susceptibility increases where there are barren land so Ranking order should be descending (Barren Class at the top i.e., 1). After Ranking, one could enter the weightage e.g on scale of 0-255, one can enter the weight of 75 to least influencing factor and proper portion of weight near to 255 for extreme. ranking weights could be computed using pairwise comparison matrix, afterward, use fuzzy membership functions to scale it from 0-255 conveniently. Surely, inflection points would be decided by the analyst himself.

I felt it appropriate to post my answer, because I found the solution and would like to share with anybody who might be facing same problem in future.

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Not the answer you're looking for? Browse other questions tagged or ask your own question.