Approach for spatial interpolation and filtering

I have several GeoTIFF images with one band (time-series of vegetation values) for one area however, taken from different orbits (positions), so I need to do either scipy.interpolate.griddata or probably use GDAL spatial interpolation for fitting all images to same coordinations (i.e. pixels of all images will have same latitude and longitude). After that, I would like to apply Savitzky-Golay filter (scipy.signal.savgol_filter) for every pixel.

What approach is best to use? I am not sure if there is a necessity to transform GeoTIFFs to DataFrame (by use of Xarray) or I can do it simply and directly with GeoTIFFs. Fitting pixels' values to the same array (of course creating lots of arrays) seems a challenging task for me.

Many thanks for any more understandable approach!