Also, there is one more change regarding the geometries: Until now, for sun zenith, sun azimuth, view zenith and view azimuth the values were taken from the center pixel. As this might be invalid in some cases, I have added that in this case the median of the data is used. I used this rather than the mean because in some cases azimuths might have very large and very small values and a large gap between these, so the mean could actually be a completely wrong solution.
Finally, this was computed twice, in the S2IdepixCloudShadowPreProcessOp and the S2IdepixCloudShadowPostProcessOp. I moved the computation to the S2IdepixCloudShadowOp and pass the results as parameters to the other ops.
Solves https://github.com/bcdev/snap-idepix/issues/44
Solves https://github.com/bcdev/snap-idepix/issues/43 . I have changed the log level and removed the logging situation so that the logging is now done by tile, not by pixel as before.
Also, there is one more change regarding the geometries: Until now, for sun zenith, sun azimuth, view zenith and view azimuth the values were taken from the center pixel. As this might be invalid in some cases, I have added that in this case the median of the data is used. I used this rather than the mean because in some cases azimuths might have very large and very small values and a large gap between these, so the mean could actually be a completely wrong solution. Finally, this was computed twice, in the S2IdepixCloudShadowPreProcessOp and the S2IdepixCloudShadowPostProcessOp. I moved the computation to the S2IdepixCloudShadowOp and pass the results as parameters to the other ops.