Hello,
We have recently had some trouble using the RasterZonalStatistics2
raster process from Geoserver and would like to share our experience. It
might be useful for someone having our same problem, or maybe be used to
improve the process's code itself.
The RasterZonalStatistics2 [1] process is distributed under the
unsupported process-raster module in geotools. As far as I know it can't
be used directly from Geoserver WPS (it's not available in the list),
but can be used to build a custom WPS process. That's how we have been
using it. Among other parameters, the process takes two main inputs: a
raster layer, and a list of features for which to compute the statistics
for.
The problem we were having is that we realized that the calculated stats
by the process were being affected by the extent of the raster layer.
Same features, different raster extent was giving different results.
After digging on the geotools code we found that the features we
provided to the process were being silently discarded when building the
ROI. The reason is that the code checks if the coverage and the features
intersect [2], but doesn't make sure that when comparing the values the
same CRS is being used. In our case the raster CRS was on degrees, and
the feature on meters, so the feature was being discarded because it
didn't intersect.
As we were using the smallest possible extent for the raster to match
the zones it took us a while to realize that the stats were being
calculated for the whole raster, and not the expected zones.
Maybe it's expected to provide all inputs in the same CRS for this kind
of process, maybe it's a general rule that I wasn't aware of. But if not
maybe one of the following improvements could be implemented:
* Document in the 'RasterZonalStatistics2' process that the 'coverage'
and 'zones' CRS must be the same.
* Add some warning log if the provided 'zones' are being discarded
because they don't interesect [2]
* Make some reprojection if both parameters don't share the CRS. Not
sure if that should be handled in the 'RasterZonalStatistics2'
process, or in the underneath 'ZonalStatistics' operation.
Thank you
[1]
https://github.com/geotools/geotools/blob/main/modules/unsupported/process-raster/src/main/java/org/geotools/process/raster/RasterZonalStatistics2.java
[2]
https://github.com/geotools/geotools/blob/main/modules/library/coverage/src/main/java/org/geotools/coverage/processing/operation/ZonalStatistics.java#L188-L190
--
*BINGEN GALARTZA IPARRAGIRRE*
magnet.coop
Santa Marina 3, Oñati (20560)
i...@magnet.coop
*www.magnet.coop*
/This e-mail and the documents attached are confidential and intended
solely for the addressee; it may also be privileged. If you receive this
communication in error, please notify the sender immediately and delete
it. If you are not the intended recipient, you are expressly prohibited
from reading, copying, distributing, disseminating or, in any other way,
using any of the information without the Magnet consent. Please note
that Internet e-mail guarantees neither the confidentiality nor the
proper receipt of the message sent, so that Magnet shall not be liable
for any damages caused. As its integrity cannot be secured on the
Internet, the Magnet liability cannot be triggered for the message
content. Although the sender endeavors to maintain a computer virus-free
network, the sender does not warrant that this transmission is
virus-free and will not be liable for any damages resulting from any
virus transmitted./_______________________________________________
GeoTools-GT2-Users mailing list
GeoTools-GT2-Users@lists.sourceforge.net
https://lists.sourceforge.net/lists/listinfo/geotools-gt2-users
