Select to view content in your preferred language

Many overlapping images as a single service

253
1
09-25-2024 02:03 PM
Labels (1)
RandyKreuziger1
Frequent Contributor

I'm a bit rusty publishing to Image Server.  In the past I've used it to publish single services of mosaic datasets containing images across a large area.  Now that I'm helping out with Image Server again I've been asked to create services for a couple dozen single images that have the exact same extent.  They are single rasters showing the results of different analyses with more planned.  Can they be mosaiced in some way as a single image service yet individually accessed?

 

Tags (2)
0 Kudos
1 Reply
TilmannSteinmetz2
Regular Contributor

Hi Randy, yes: I think the easiest way to achieve that would be to create a new Mosaic Dataset in a FileGeodatabase. Then add all those rasters to the Mosaic Dataset, which can then be published to Image Server as a single Image Service. The MDS can get any number of additional attributes added to its attribute table, which can be used to add, e.g. time and date for different image dates for overlapping rasters.

The image service gives you a single handle for all rasters by being able to return results depending on input parameters, such as certain attribute values. In other words, the image service is able to 'execute' a query to the Mosaic Dataset (which behaves like a searchable catalogue) and return matching results to the user.

Once you've published the image service, just pop it in a web map in ArcGIS Online. There is one InstantApp template which lends itself well as a simple UI for making that kind of request and displaying one of the images/rasters at a time.

(If you need to query _all_ the rasters at a location, e.g. for a timeseries plot of values for a pixel, you would create the MDS as above, add the rasters, calculate and required field values, such as StdTime, build multidimensional info and then export the whole thing into another format: CRF, or cloud-raster-format. There's a setting to calculate the multidimensional transpose, which makes querying pixels through the entire stack of rasters extremely fast). The resulting T-CRF cache structure ('tile pile') can again be published as an image service.

0 Kudos