diff --git a/README.md b/README.md index 60f9088549090af85b393792a6ece1adb95cbe4a..e81d9a5a4ab35b23613703de88c3f1162522649e 100644 --- a/README.md +++ b/README.md @@ -5,7 +5,7 @@ This module provides one application dedicated to clear cuts detection Clear cuts detection ---------------------------- -This application implements the clear cuts detection pipeline, developed by IRSTEA (Kenji Ose, Michel Deshayes) +This application implements the clear cuts detection pipeline, developed by IRSTEA (Kenji Ose, Michel Deshayes, Rémi Cresson) ###Method### From a pair of images, (each one aquired at a different date), and an optional forest mask, dNDVI (the difference of the normalized difference vegetation index) is computed. Then, mean and standard deviation are both computed from this resulting image. The last step is the multiple thresholding of dNDVI (based on mean and standard deviation) in order to produce a map of detected clear cuts confidence. The actual values can be 0,1,2 or 3 (A value equals to 3 beeing the most high NDVI difference range, i.e. the most realistic clear cut detection).