nav[aria-label="Primary Navigation"] { padding: 0; & ul { list-style: none; width: 100%; display: flex; flex-direction: row; justify-content: start; align-items: start; gap: 30px; padding: 0; & li { margin: 0; } & ul li { list-style: none; } } }

Generate Mosiac Dataset from High Resolution Imagery

TrentReeder

I am looking for some general recommendations on which set of files to use when generating a mosaic dataset. We have 3 inch, county wide imagery in two different formats...

1 folder containing 5 MrSID files ranging from 7gb to 36 gb in size
1 folder containing 11,709 TIF files that around 155 mb each in size

I originally created a mosaic dataset using scripts and workflow from Managing Preprocessed Orthophotos. Going through the workflow would have created more than 22,000 overviews with the MrSID files. Not sure if this is an excessive amount of overviews or not. If I do not Force Overview Tiles, only 6 overviews are created.

I am thinking that maybe using the the smaller TIF files may be the way to go. I wanted to ask before I went too far down the rabbit hole.

The end goal will be serving up the imagery as an image services basemap. We are using 10.7.1 Enterprise with the Image Server extension

Thanks for your help!

Find more posts tagged with

Visualization

Mosaic datasets

Image Services

Accepted answers

PeterBecker

You could create the mosaic data and image service from both. Best if you open a TIF file and MrSID file of the same area and review the difference in the image quality. Also check the number of bands. Are they both 3band or is the TIF 4band (IR)? The MrSID Files are highly compressed (loosy) and you will see some difference in the data values. These differences should be small and if acceptable for your applications then I would recommend using the MrSID files. The MrSID files include pyramids, which means that few overview are required. You can go with the default numbers of overviews, else set the pixel size of the overviews to about 10x the pixel size of the source. The overviews will still be small in comparison with the source.

If you want the better quality you will need to go with the TIF files. It is likely though that the TIF files you have are not optimized and do not have pyramids, hence will be slower to access and require more overviews. I would recommend that you use OptimizeRasters to convert the TIF files to TileTIF, MRF or COG. As part of this optimization you can choose the compression and for lossless and or lossy. Again if using lossy you need to determine the quality level. The size should come down to about 1/8 of the source data with negligible quality loss. OptimizeRasters will also build pyramids for the data so again the number of overviews required would be low. Let ArcGIS determine the default or set to about 8x pixel size. As George mentioned, do look at the Imagery Worfklows, specifically the section on Imagery Formats and Performance

All comments

George_Thompson

I always start here: ArcGIS Imagery Workflows | Documentation

Here is some specific information related to imagery management and the mosiac dataset: Using Mosaic Datasets to Manage Imagery—Imagery Workflows | Documentation

PeterBecker

TrentReeder

Thanks, Peter! This is exactly what I was looking for. I started building our imagery server last month and have been reading and rereading the Imagery Workflows documentation on the various building blocks. The documentation has been instrumental in getting things up and running.

I should have added further info...MrSID is 4band and the TIF files are 3band with no pyramids.

I will look into OptimizeRasters. For some reason I thought those workflows were more for working with cloud storage data.

Thanks again!

Join the Conversation

Signed in members can post, follow updates, and more. New here? Register a free account.

Quick Links

Community Help
Find useful guides, FAQs, and documents to help you navigate and make the most of Esri Community.