I have ndvi values from series of points obtained from time series satellite images. How can i detect anomally from these values? Thanks
Hello, the answer depends on what kind of anomaly you want to reveal (spatial, temporal, spatio-temporal). I would not mix up NDVI for different areas (at least I would not mix up negative NDVIs for water with positive NDVIs for soil). If we consider a series of maps over a long period (e.g. over a year) then for each lat/lon bin one should build the average value and subtract it from the corresponding lat/lon bin of each of the frames. The series of these new frames will show how NDVI anomaly develops with time.
Artem G. Feofilov::: Thank you Sir, I actually want to detect spatio-temporal anomally. Thanks for the highlight about mixing the ndvi for different areas it solved one of my problems.
Pls i am not clear about the lat/lon bin, do you mean rectangular bins for different areas temporally?
Milad Janalipour;;; Thank You Sir, i tried the RXD but it requires image files as input, and the inputs i have at hand are values (Extracted figures), pls is there another way to input them?
> Pls i am not clear about the lat/lon bin, do you mean rectangular bins
> for different areas temporally?
for simplicity, let's consider one year of data for 1 deg lat x 1 deg lon bin at lat=0 and lon=0: we have a sequence of 365 numbers and we can subtract their average from this sequence. This way we obtain a temporal NDVI anomaly I was talking about. If we repeat the procedure for other lat/lon bins, we will have a sequence of 365 maps with anomaly values for each pixel, and then it's up to you how to use this information :)
Ok thanks a lot for the clarification Sir. Please is it possible to get any reference or link for this method to cite in my work? Thank you
you are welcome.
However, I can't give you any reference because the procedure I described above comes directly from the definition of the anomaly. Different researchers use it for different purposes in different fields, but the approach remains close to what is written.
- Badges
- Science topic
- Similar topics
- Cell Biology
- Methods