Create and use a mosaic dataset

In this tutorial, you'll focus on the management and storage of large volumes of imagery in ArcGIS Pro. As an imagery analyst for the Upper Austria government, you have received a collection of aerial orthophotos that you must manage and share effectively with stakeholders. You will first explore the challenges of working with multiple images individually. You will then create a mosaic dataset that will allow you to work with all the images as a single set, turning them into useful information products for both visualization and analysis. Finally, you'll enhance the mosaic dataset visually and perform a short analysis.

This tutorial was last tested on July 16, 2026, using ArcGIS Pro 3.7 . If you're using a different version of ArcGIS Pro, you may encounter different functionality and results.

Requirements

• ArcGIS organizational account with a Publisher, Facilitator, or Administrator role (see options for software access)
• ArcGIS Pro (see options for software access)

Work with multiple individual images

Large collections of imagery and raster data are being collected and assimilated at a rapid pace, and these datasets have significant value when the information inside them is shared and disseminated. ArcGIS Pro provides extensive image management capabilities and is used by organizations in a wide range of industries to manage large imagery holdings, making them accessible and turning them into useful information products for both visualization and analysis. Managing imagery and raster data efficiently and correctly is key to ensuring accessibility. For this, ArcGIS Pro uses mosaic datasets, a type of geodatabase structure, to manage imagery. In addition, ArcGIS Pro includes all the technology and associated tools needed to build and maintain mosaic datasets.

In this first part of the tutorial, you will explore the challenges of working with multiple images individually.

Note:

This tutorial doesn't address the basics of imagery. If you are new to this topic, consider first doing the introductory tutorial Get started with imagery.

Set up the project

First, you'll download the imagery data you'll use in this tutorial and set up anArcGIS Pro project.

1. Download the Orthophotos.zip file.

   The compressed .zip file downloads to your computer. It contains 73 aerial orthophoto images that cover the region in and around the historic town of Hallstatt, Austria.

   Note:

   This set of images is large and may take some time to download.

   Most browsers download to your computer's Downloads folder by default.

   You'll uncompress the file to a folder named HallstattProject on your C: drive.

2. In Microsoft File Explorer, go to your Download folder, right-click the Orthophotos.zip file, and choose Extract All.

   

3. For Files will be extracted to this folder, type C:\HallstattProject and click Extract.

   

4. After the extraction is complete, go to C:\HallstattProject\OrthoPhotos, and verify it contains a large number of image files, with names such as 4727-08.jp2.

   

   Note:

   The data used in this tutorial comes from the State of Upper Austria (Land Oberoesterreich Open Data), which provides data for their region under a Creative Commons Attribution 4.0 Austria license. If you have questions about this data, use the following contact information:

   Office of Upper Austria State Government,
   Directorate Presidium, Department Presidium,
   Landhausplatz 1, 4021 Linz, Austria
   Telephone [phone]

   A list of available data is located on the State of Upper Austria's site.

   You will now create an ArcGIS Pro project to work with this data.

5. Start ArcGIS Pro. If prompted, sign in using your licensed ArcGIS organizational account.

   Note:

   If you don't have access to ArcGIS Pro or an ArcGIS account (for ArcGIS Online or ArcGIS Enterprise), see options for software access.

6. Under New Project, click Map.

   

7. In the Create a New Project window, for Name, delete the default project name and type HallstattProject.

   By default, projects are saved in a new folder; however, you'll save this project to the same folder where you stored your imagery data.

8. For Location, browse to C:\, select HallstattProject, and click OK.
9. Uncheck Create a new folder for this project and click OK.

   

   The project is created with a default map, that is set to the world extent and contains only the World Topographic Map basemap.

   

Display two individual images

You will now learn about the challenges of working with multiple images individually. Even though adjacent images may appear as a single image when displayed in a map, they are separate layers. Working with the individual layers is challenging when project requirements call for the application of any kind of enhancement or analysis because each layer must be handled separately. You will start exploring the set of images. They are accessible through the Catalog pane

1. On the ribbon, click the View tab. In the Windows group, click Catalog Pane.

   

2. In the Catalog pane, expand Folders and HallstattProject.

   

   The HallstattProject folder contains the HallstattProject.gdb geodatabase and other items created with the project. It also contains the OrthoPhotos folder that you will now explore.

3. Expand the OrthoPhotos folder.

   

   It lists the set of 73 images in a .jp2 file format.

   Note:

   A .jp2 file is a compressed bitmap image created using JPEG 2000 (JP2) Core Coding. It supports color bit depth and image metadata and may be compressed with lossy or lossless compression. It is typically used for storing large digital photos and images.

4. Expand the first file, named 4727-08.jp2.

   

   The image consists of four spectral bands. While they are not named as such, bands 1, 2, 3, and 4 represent the red, green, blue, and infrared spectral bands, respectively. All 73 images in the set have four bands.

   Note:

   Learn more about spectral bands in the tutorial Explore imagery: Spectral resolution.

5. In the OrthoPhotos folder, collapse 4727-08.jp2.

   You will add the first two images to the map.

6. Press the Ctrl key and select the images 4727-08.jp2 and 4727-16.jp2. Right-click the selection and choose Add To Current Map.

   

7. When prompted to Calculate statistics, check the Always use this choice box, and click Yes.

   After the statistics have been calculated, the two images are added to the map and they are listed in the Content pane.

   

   On the map, the two images are adjacent to each other.

   

8. On the Map, zoom in with the mouse wheel to the junction of the 4727-08.jp2 and 4727-16.jp2 layers.

   The two images are displayed independently of each other, so you can see that there is an abrupt change from one to the other, as the color and brightness appearance is different for each one.

   

9. In the Contents page, if necessary, click 4727-08.jp2 to select it.

   

10. On the ribbon, click the Raster Layer tab. In the Enhancement group, change the Brightness level to 10.

    

    The lower image updates to appear brighter.

    

    Note:

    This type of enhancement is only applied to the image display, and doesn't affect the underlying image file.

11. In the Enhancement group, experiment with changing the Brightness, Contrast, and Gamma value for one image or the other. Observe how the display is affected.

    You can see that the display of these images is handled completely separately. And these are only two of the 73 images in your set. The more images you have, the harder it will be to display them with an appearance that looks seamless. Running an analysis workflow on each of these 73 images individually would be similarly challenging. In the rest of the tutorial, you will learn how solve that problem by creating a mosaic dataset that contains your entire image collection--a dataset that will behave as if it were a single very large image. This is a powerful approach to manage an image collection.

    You don't need the two individual images any longer, so you'll remove them from the map.

12. In the Contents pane, press the Ctrl key and select 4727-08.jp2 and 4727-16.jp2. Right-click the selection and choose Remove.

    You'll save your project.

13. In the Quick Access Toolbar, click the Save Project button.

    

You added separate images to a map and explored enhancing them. You learned that working with many images individually poses challenges and it not the best approach.

Create a mosaic dataset

You will now create and populate a mosaic dataset to efficiently manage and display your aerial image collection. Then, you'll adjust its properties.

A mosaic dataset is a well-defined geodatabase structure optimized for working with large collections of imagery and rasters. Mosaic datasets are stored in either a file geodatabase or an enterprise geodatabase. The imagery and raster data do not need to reside in the database. Most organizations store their imagery as files on disk, enterprise, or cloud storage. A single mosaic dataset can reference millions of images and make them appear as a single virtual dataset. It also enable quick access to any of the individual image it contains. With mosaic datasets, the large volume of pixel data (contained in the imagery and rasters) are not loaded into the database and are instead referenced. The metadata about the data sources, as well as information on how to process the imagery into different products, is stored in the mosaic dataset. When a request for imagery is made, the mosaic dataset is used to determine what images are required and what processing needs to be applied. Only the required imagery is read, processed, and returned.

Create an empty mosaic dataset and populate it

You'll create an empty mosaic dataset and populate it with your 73 images.

1. If necessary, in the Catalog pane, expand Folders and HallstattProject. Collapse the OrthoPhotos folder.

   

2. In the HallstattProject folder, right-click HallstattProject.gdb, point to New, and choose Mosaic Dataset.

   

   The Create Mosaic Dataset geoprocessing tool appears.

3. In the Create Mosaic Dataset pane, for Mosaic Dataset Name, type Hallstatt.
4. For Coordinate System, choose Current Map [Map].

   The value updates to MGI_Austria_GK_Central

   Note:

   As you displayed some of the images earlier, the map was automatically set to their coordinate system, which is MGI_Austria_GK_Central. You'll use that same coordinate system for the rest of the tutorial, as it is well adapted to the Austria geographic area.

   In general, it is best practice to set the mosaic dataset to the same coordinate system as the images it will contain.

   

5. Click Run.

   The tool creates a mosaic dataset named Hallstatt in the HallstattProject geodatabase. In addition, it adds that new dataset to the Contents pane of the map.

   

   The mosaic dataset is currently empty. Next, you'll populate it with your 73 images.

6. At the bottom of the Geoprocessing pane, click the Catalog tab to activate it.

   

7. In the Catalog pane, expand the HallstattProject.gbd geodatabase, right-click the Hallstatt mosaic dataset, and choose Add Rasters.

   

   The Add Rasters To Mosaic Dataset tool appears.

8. For Raster Type, confirm the value is set to Raster Dataset.

   Since the imagery you are working with is simple imagery without any metadata, you'll use the default Raster Dataset raster type.

   Next, you'll specify the input data type and location. You will point to the entire folder where the 73 images are located.

9. For Input Data, expand the drop-down list and choose Folder (Optional).
10. Click the Browse button. Browse to Folders, HallstattProject, select the OrthoPhotos folder and click OK.

    

11. Expand the Raster Processing group option. For Minimum Rows or Columns, type 10.

    

    By specifying a value of 10, you are allowing the mosaic dataset access to all the pyramids from the source images that have 10 or more rows or columns.

12. Click Run.

    When the process is complete, you'll zoom in to the mosaic dataset extent.

13. In the Contents pane, right-click Hallstatt and click Zoom To Layer.

    

    The mosaic dataset appears on the map.

    

    You see a footprint grid (in green) that represents the location of every added images. Some of the images appear in the grid but not all, you'll remedy that issue later in the tutorial.

14. Press the Ctrl+S shortcut to save the project.

Explore the mosaic dataset

Next, you'll explore the mosaic's layers and attribute table.

1. In the Contents pane, examine the Hallstatt mosaic dataset.

   It is a group layer containing three sublayers:

   • Boundary
   • Footprint
   • Image

   

   You'll review them one by one.

2. In the Contents pane, check the box next to the Boundary sublayer to turn it on, and uncheck the Image and Footprint sublayers to turn them off.

   

   The map update to show only the Boundary sublayer. It is a feature class that represents the combined footprints of all the images added to the mosaic dataset.

   

3. In the Contents pane, turn off the Boundary sublayer and turn on the Footprint sublayer.

   The Footprint sublayer is a feature class that represents the footprint of each image added to the mosaic dataset.

   

   In the Footprint layer, you can see the metadata collected about every image. You'll open the attribute table to review that information.

4. In the Contents pane, right-click the Footprint sublayer and choose Attribute table.

   

   The attribute table appears.

   

   Every line of the table represents an image. If you scroll down, as expected, there are 73 rows. The fields tell you various information about each image, including the following:

   • Its name (Name) -- so that the specific image can be found in the folder where it is stored.
   • The coordinates of its center (CenterX and CenterY) --so that the image can be correctly positioned on the map.
   • Information about its pixel size (MinPS, MaxPS, LowPS, and HighPS) --so that the image can be displayed optimally at different scales

   All this metadata ensures that the mosaic dataset can properly display large collections of images.

   Note:

   Learn more about these fields in the Mosaic dataset attribute table documentation page.

   In this case, the collection of aerial images didn't contain much metadata originally. If you create a mosaic dataset with images with metadata and you specify the appropriate raster type, the attribute table will be populated with far more attribute fields with values derived from the original metadata.

5. Close the attribute table.

   

6. In the Contents pane, turn on the Image sublayer.

   

   This layer represents the images themselves, displayed as a seamless mosaic. Currently not all 73 images are displaying. Next, you'll remedy this, as well as adjust some other properties of the mosaic dataset.

Set mosaic properties

You'll adjust the properties of the mosaic dataset. Currently only the first 20 images are displaying, because that's the default. You will choose a higher number.

1. In the Catalog pane, under the HallstattProject.gdb geodatabase, right-click the Hallstatt mosaic dataset and click Properties.

   

2. In the Mosaic Dataset Properties window, click the Defaults tab, and, if necessary, expand the Image Properties section.

   

   The Defaults tab allows you to control how the mosaic dataset is displayed and handled by default.

3. In the Image Properties section, for Maximum Number of Rasters Per Mosaic, type 100 and press Enter.

   This property sets how many images are displayed in the mosaic dataset. Since your mosaic dataset consists of 73 images, setting the maximum value to 100 ensures all source images will be displayed.

   

4. In the Mosaic Dataset Properties pane, click OK.

   All 73 images appear on the map, along with the footprint.

   

5. In the Contents pane, turn off the Footprint sublayer.

   The map updates and the mosaic dataset now displays as a seamless image collection, appearing like a single large image. Note that are no difference from one image to the other, in terms of brightness, contrast, or other display properties, since the whole dataset is handled as a whole.

   

6. In the Contents pane, under Image, right-click Red: Band_1 to display all the spectral bands available.

   The bands are named Band_1, Band_2, Band_3, and Band_4, which is not very informative. You know they correspond to the colors Red, Green, Blue, and Infrared, so you would like them to be named that way. You will also change that in the mosaic dataset properties.

7. In the Catalog pane, under the HallstattProject.gdb geodatabase, right-click the Hallstatt mosaic dataset, and click Properties.

   

8. In the Mosaic Dataset Properties window, click the General tab, and, if necessary, expand the Raster Information section.

   

   In the General tab, you find information regarding the mosaic dataset such as source information, location, and whether it has statistics or not.

9. In the Raster Information section, locate Product Definition and click the Edit button next to NONE.

   

   The Product Definition window appears. This property enables you to specify the type of imagery the mosaic dataset represents, resulting in the setting of several preset default values, including the band names. For now, Product Definition is set to NONE.

10. In the Product Definition window, expand the dropdown list, and choose NATURAL_COLOR_RGBI.

    Note:

    RGBI stands for Red, Green, Blue, and Infrared.

    The Product Band Definitions settings update, listing the bands with their correct color names and the minimum and maximum wavelengths for each band.

    

    Note:

    The wavelength ranges indicated represent common averages for these color bands.

11. Click OK. Click OK again.
12. In the Contents pane, under the Image sublayer, right-click Red.

    The list of the four bands appear, displaying the new color names and wavelength ranges.

    

    Currently, the Red band displays through the Red channel, the Green band, through the Green channel, and the Blue band, through the Blue channel. This forms the natural color band combination that approximates the way the human eye sees the landscape.

    Note:

    Learn more about how the spectral bands are displayed through the RGB channels in the tutorial Explore imagery: Spectral resolution.

13. Press the Ctrl+S shortcut to save the project.

You created a mosaic dataset and populated it with your aerial image collection; then you adjusted the mosaic's properties.

Use a mosaic dataset as a dynamic image

In this last part of the tutorial, you will work with the mosaic dataset you created. You will enhance it visually and perform a short analysis.

Enhance the mosaic dataset visually

First, you'll enhance the dataset visually.

1. In the Contents pane, confirm the Image sublayer is turned on and the Footprint and Boundary sublayers are turned off.

   

   One important type of visual enhancement is to apply a stretch. A stretch improves the appearance of an imagery dataset by spreading the pixel values from the darker to the lighter tones to obtain a more vivid rendition.

2. In the Contents pane, select the Hallstatt mosaic layer.
3. On the ribbon, on the Mosaic Layer tab, in the Rendering group, click the Stretch Type drop-down menu.

   Currently there is no stretch applied to the mosaic dataset.

   

   To apply stretches and other visual enhancements on the mosaic dataset, you must first build statistics, which summarize the pixel values for the entire dataset. When a mosaic dataset is created and images are added to it, statistics are not automatically generated, since calculating statistics for many images may be time consuming. Once the mosaic dataset is built, it is advisable to calculate statistics to enable visual enhancements and improve performance. You will do that now.

4. In the Catalog pane, under the HallstattProject.gdb geodatabase, right-click the Hallstatt mosaic dataset, point to Enhance, and choose Calculate Statistics.

   

5. In the Calculate Statistics tool, accept all the defaults and click Run:

   

   Calculating statistics for this mosaic dataset may take a few minutes to complete.

6. When the process is complete, in the Catalog pane, right-click the Hallstatt mosaic dataset, and click Properties. In the Mosaic Dataset Properties window, click the General tab, expand the Statistics section.

   

   The mosaic dataset statistics have been computed and now include pixel value information for each band, such as Minimum, Maximum, Mean, and Standard Deviation values.

7. Close the Mosaic Dataset Properties window.

   After the statistics were computed, the display of the entire mosaic dataset was updated on the map and a default stretch was automatically applied to it. For convenience, you can see the change in the following image examples:

   

   Compared to the image on the left that has no stretch, the image on the right, with the stretch has crisper, move vivid colors. In particular, the snowy areas in the southern side are brighter, and the lake in the northern side is darker.

8. In the Contents pane, ensure the Hallstatt mosaic layer is still selected.
9. On the ribbon, on the Mosaic Layer tab, click the Stretch Type drop-down menu again.

   

   The Stretch Type is now set to Percent Clip.

10. On your own, experiment with various stretch types to see how the affect the imagery rendering differently. When you are finished, change the Stretch Type back to Percent Clip, as this is an excellent default for this dataset.

    Note:

    Learn more about the Percent Clip stretch and other types of stretches.

    While the dataset rendering has improved with the stretch, you still would like it to have it display a bit brighter. You will do that now.

11. On the ribbon, on the Mosaic Layer tab, in the Enhancement group, set the Brightness value to 5 and press Enter.

    

    The entire dataset updates to be slightly brighter.

    

12. Optionally, experiment with different values for Brightness, Contrast, and Gamma. When finished, click on the Reset buttons next to each slider to go back to the original values, and set Brightness to 5 again.

    

    Note:

    Any change to the appearance of a layer, such as stretch, brightness, contrast, and gamma, is only for visual display purposes and will not actually change the source data.

13. Press Ctrl+S to save the project.

    Note:

    Optionally, to accelerate the display of your dataset, you can also build overviews. These are reduced resolution (or miniature) images that represent different parts of your mosaic dataset. In the Catalog pane, right-click the Hallstatt mosaic dataset, point to Optimize, and choose Build Overviews. Accept the defaults and click Run.

Perform an analysis on the mosaic dataset

You will now perform a short analysis on your mosaic dataset. You want to highlight the presence of healthy vegetation in the image. You will do that with the Normalized Difference Vegetation Index (NDVI), a formula that computes a ratio between the near infrared and red bands.

Note:

Learn more about NDVI.

1. On the ribbon, on the Imagery tab, in the Analysis group, click the Raster Functions button.

   

   The Raster Functions pane appears and shows various functions that you can use to perform analysis on your imagery.

2. In the Raster Functions pane, under Analysis, click the NDVI Colorized function.

   

   NDVI Colorized computes the NDVI index, and it then applies a color symbology, so that vegetation appears in green and non-vegetation, in orange tones.

3. In the NDVI Colorized Properties pane, set the following parameters:
   • For Raster, choose Hallstatt.
   • For Visible Band ID, choose 1.
   • For Infrared Band ID, choose 4.

   In your imagery, bands 1 and 4 correspond to the red and near infrared bands respectively.

4. Leave the other parameters unchanged and click Create new layer.

   

   A new NDVI Colorized_Hallstatt layer appears.

   

   You will compare it to the mosaic dataset in natural color using the Swipe tool.

5. In the Contents pane, confirm the NDVI Colorized_Hallstatt layer is selected.

   

6. On the ribbon, on the Raster Layer tab, in the Compare group, click Swipe.

   

7. Drag back and forth from top to bottom on the map to reveal the natural color imagery under the NDVI Colorized_Hallstatt layer.

   On the new layer, areas of healthy vegetation appear in green and areas low in vegetation in cream color. Water bodies appear in orange.

   

   This layer is a great way to highlight healthy vegetation in your imagery.

8. On the ribbon, on the Map tab, in the Navigate group, click the Explore button to exit the swipe mode.

   

   This was a basic one-step analysis example, however you could also combine several of raster functions and other geoprocessing tools to create much more complex analyses. To explore more imagery analysis examples, see the following tutorials:

   • Assess hail damage in cornfields with satellite imagery
   • Classify land cover to measure shrinking lakes
   • Calculate impervious surfaces from spectral imagery
   • Classify land cover to measure shrinking lakes
   • Extract high-resolution land cover with geospatial AI

   The workflows in those tutorials are performed on a single images, but they could be performed on a large mosaic dataset as well, without having to modify the workflows in any ways.

Add analysis to a mosaic dataset

Raster functions can be applied quickly to very large mosaic datasets. However, when applying one or several of them to a dataset, the layer output is dynamically computed and not saved on disk. One way to give reliable access to the raster function-based analyses you develop is to attach them to the dataset. You will do that now with the NDVI analysis you just performed. As a result, anyone using your mosaic dataset will be able to apply that analysis whenever they wish. First, you'll save the analysis.

1. In the Contents pane, right-click the NDVI Colorized_Hallstatt layer and choose Save Function Chain.

   

   The Raster Function Editor pane appears.

2. In the Raster Function Editor pane, click the Save button.

   

3. In the Save window, set the following parameters:
   • For Name, type NDVI Colorized.
   • For Category, choose Project.
   • For Description, type Visualize vegetation in bright green, low vegetation or shadow areas in light tones, and water in orange.
   • Click OK

   

4. Click OK.

   You have now saved the analysis as a raster function template that can be associated to the mosaic dataset as a processing template.

5. Close the Raster Function Template 1 pane.

   

6. In the Contents pane, right-click the NDVI Colorized_Hallstatt layer and choose Remove, as you don't need it any longer.

   

7. In the Catalog pane, right-click the Hallstatt mosaic dataset and choose Manage Processing Templates.

   

8. In the Manage Processing Templates pane, click the Import button.

   

9. In the Select Processing Templates window, click Raster Functions and browse to Project and Project1. Select NDVI Colorized.rft.xml and click OK.

   

10. Click OK.

    The Manage Processing Templates pane updates and displays the attached NDVI Colorized raster function.

    

    Now that you have attached the processing chain to the mosaic dataset, you or other users can apply the analysis at any time.

11. In the Contents pane, select the Hallstatt layer.
12. On the ribbon, click the Data tab. In the Processing group, click the Processing Templates drop-down menu, and choose NDVI Colorized.

    

    The map updates to show the results of the raster function template applied to the layer.

13. Click the Processing Templates drop-down menu again and choose None.

    Your display changes back to the original imagery rendering.

    

    If you develop other analyses in the future, you could similarly associate them to the mosaic dataset as additional processing templates.

14. Press Ctrl+S to save the project.

In this tutorial, in your role as an imagery analyst for the Upper Austria government, you received a collection of aerial orthophotos that you needed to manage and share effectively with stakeholders. You first explored the challenges of working with multiple images individually. Then you created a mosaic dataset to allow you to work with all the images as a single set, making them accessible and turning them into useful information products for both visualization and analysis. Finally, you enhanced the mosaic dataset visually, performed a short analysis with a raster function, and associated that analysis to the dataset as a processing template.

Acknowledgements

• Orthophoto images provided by Land Oberoesterreich Open Data. Data source: Land Upper Austria - data.ooe.gv.at.
• Multispectral Landsat imagery layer sources: Esri, USGS, NASA, and AWS.
• World Topographic Map sources: Esri, TomTom, Garmin, FAO, NOAA, USGS, OpenStreetMap contributors, and the GIS User Community