2015-05-03 WV3 strips for Sindhupalchok and Kavrepalanchok districts

[cgiovando]

There are three strips of Worldview 3 collected on May 3, 2015 which cover a large area of the affected districts of Sindhupalchok and Kavrepalanchok. Same strips are also referenced in #16, but because of the urgency, we are going to make it a separate request. This imagery would fill the gap in IDP camps mapping between tasks 1030 and 1046.

This is also the same area hit again by the May-12 M7.3 earthquake, so there also urgent requests for mapping other elements such as potential helicopter landing zones.

Individual source scenes are available on HDDS, but are not ortho corrected, pansharpened and coregistered. The actual strip mosaics seem to be available through the DG catalog, but not found on their Nepal S3 bucket.

For reference, here are the links to the catalog:

https://browse.digitalglobe.com/imagefinder/showBrowseMetadata?catalogId=104001000A499E00 https://browse.digitalglobe.com/imagefinder/showBrowseMetadata?catalogId=104001000B55EA00 https://browse.digitalglobe.com/imagefinder/showBrowseMetadata?catalogId=104001000A702C00

We need urgent access to the processed mosaiced strips. We will then have capacity to create a tile service for posting to HOT for mapping.

[d3netxer]

My understanding is that for right now, these three strips are available on the re-caching DG service, but they are being requested in case new overlapping images obscure these three.

Yesterday, I believe two different people were looking into it they could provide these three strips (Google & GeoCenter). I do not know what their status is right now.

Since it seems this request is urgent and has been coordinated, I suggest the request should be sent by e-mail to the image-coord listserv.

[cgiovando]

The OSM-fr proxy cache is a mix of multiple days collects, not ideal for mapping the situation at certain point in time. The tiles levels are also not consistent (e.g. you may see something at Z=15 and then something different at Z=16).

Google (@jlivni) confirmed that they don't have these strips. Not sure about GeoCenter, do you have a contact there to confirm? Have pinged @kevinbullock but waiting for reply. I know Chad@USAID is trying his channel as well.

I emailed imagery-coord right after posting here, waiting on any reaction there :-)

[d3netxer]

Chad@USAID is from GeoCenter. I should see him later today and I'll ask him. Thanks

[kevinbullock]

These images are available through DigitalGlobe's WMS service (access details below). I did make this request to our team providing imagery through S3 as well. They placed the order yesterday, but I do not have a status update. Perhaps check our S3. The images we are providing are orthorectified and pan-sharpened, but are not mosaicked together.

http://services.digitalglobe.com Username: nepal Password: forcrisis​

[cgiovando]

Thanks @kevinbullock . The problem with the WMS is that it returns a mix of different collects all overlapped to each other. The most recent imagery is more cloudy than May 3rd's, so unless there's TIME support in the WMS service, it's going to be hard to use it.

Anyway, I let @d3netxer take the lead in direct requests to DG since HIU already has an established agreement. Thanks for all your support so far, we really appreciate. We'll keep an eye on the S3 bucket in the meantime.

[kevinbullock]

@cgiovando got it, thanks. Time is indeed a variable that can be filtered through a CQL query in WMS. (http://docs.geoserver.org/stable/en/user/tutorials/cql/cql_tutorial.html) . Additionally, a specific feature (image) can be requested via WCS request. The featureID is in the image metadata. (http://docs.geoserver.org/latest/en/user/services/wcs/reference.html).

More than happy to coordinate with HIU. My only gripe is that we'd be duplicating work in terms of imagery processing, but, HIU will be able to give this more touch. Please do keep an eye on S3 as well, my point of contact, Jordan Winkler is unfortunately traveling today, else I'd include him in this thread.

Regards, Kevin

[cgiovando]

Thanks @kevinbullock, trying to pull out the correct URL for a filtered 2015-05-03 example, but not getting anything with https://services.digitalglobe.com/mapservice/wmsaccess?connectid=71ae32ec-f6df-4ed7-b707-2aa016610679&CQL_FILTER=TIME%20=%20'2015-05-03'

I also tried the more standard WMS TIME parameter, but it doesn't seem to make a difference:https://services.digitalglobe.com/mapservice/wmsaccess?connectid=71ae32ec-f6df-4ed7-b707-2aa016610679&TIME=2015-05-03

Any direct example you can point me to? And where do I find the WCS endopoint? (I only see WMS, WMTS and WFS).

I understand your concerns with duplicating work. When mentioning "processing" I'm referring to just setting up a tile map service, especially if the imagery is already orthocorrected and coregistered. There have also been moments when the WMS was overloaded an setting up a separate TMS at HIU would make things more scalable for HOT and less traffic to handle for you. But again, I'll let @d3netxer coordinate with your team to avoid overlaps.

[EdHankin]

Hi @cgiovando, Please see https://www.digitalglobe.com/connectonline for our OGC dev guides. These should be helpful for CQL and other common formatting questions. Please note that the URLs you posted are missing some required parameters, again these are specified in the dev guides.

This is how i would suggest filtering for date and source to bring these images to the top:

&COVERAGE_CQL_FILTER=(acquisitionDate<%272015-05-03 00:00:00%27)%20AND%20(acquisitionDate>%272015-05-04 00:00:00%27)%20AND%20(source=%27WV03%27)

Also, you may consider setting a cloud cover limit, which may be more useful for a layer that is not specific to a single date.

As with anything, sometimes these filters take some iteration to accomplish the desired outcome.

Regarding WCS, for this public access account it has not been enabled.

Regards,

Ed Hankin DigitalGlobe Technical Support cloudservices@digitalglobe.com