Closed FlashThunder closed 3 years ago
Having a look at the estimated variogram parameters:
>>> print(uk.variogram_model_parameters)
[1.34819582e-11 6.83151812e+06]
This reveals that a pure nugget model was estimated (flat slope, high nugget). Therefore the constant output.
Tinkering around with your input data reveals, that estimating and fitting a variogram is rather hard, since spatial correlation is hard to find. so maybe another routine for interpolating your data could be better.
Oh ok, I understand that my data doesn't fit to interpolate that way. But could You give me a hint on how to fix it? And be able to use it to estimate the missing data, as well as "correct" values? The model seems to work great on examples. It is hard to make a custom variogram, as data differs. I would like to use it to fill missing data (clouds etc) on Sentinel 2 images. I mean having values from different days.
Why doesn't it even try to interpolate the given values? just returns single value for every y: 2356.24193548.