I have several months worth of NDVI images for an area. When I check the values, there seems to be some deviation. I am using python GDAL to attempt to calculate possible areas of concern for inspection.

For Example:

Month     X       Y         V 
Jan      55.45  117.25   0.058742
Feb      55.45  117.25   0.0467307
Mar      55.45  117.25   0.153668

And here would be the area

enter image description here

When I compare the NDVI images, some of them seem off. So somehow I need to normalize the datas indexes to find the changes.

I'm kind of stuck and looking for some advice. Perhaps generate VCI's for each image and compare the percentage of change over the area? Each images indexes are correct with itself, but not always relative to the others covering the same area, but I do believe this will work because the min and max values are relative to that image.


But the indices are relative only to that image, or so it seems. So the variations within that image are relative to that image only. When I start subtracting to find the changes I end up with some areas that when inspected have not seemed to change.

My objective is to find changes with significant vegetation death and go inspect the cause.

  • 1
    What is your ultimatel objective? I would expect NDVI values to change from month to month. In your area if interest, would there be more healthy, green vegatation in March over Jan or Feb?
    – Aaron
    Jan 15, 2020 at 14:58
  • Please see @JeffreyEvans response recommending the Theil-Sen approach to derive the slope(s) and intercept(s): researchgate.net/post/…
    – Aaron
    Jan 15, 2020 at 15:16
  • Could you explain how to do that in python? his example is in R
    – Messak
    Jan 15, 2020 at 15:22


Your Answer

By clicking “Post Your Answer”, you agree to our terms of service and acknowledge you have read our privacy policy.

Browse other questions tagged or ask your own question.