Imagery in Action

Exercise
Use satellite data to map a land
cover type
Section 2 Exercise 2
August 17, 2021
 Imagery in Action

Use satellite data to map a land cover type

Time to complete
40 minutes

Introduction
Imagery layers in ArcGIS Pro can be used from online and file-based sources for analysis and
displayed at the same time in the same map. Depending on the source of the imagery layer,
the functionality and visualization options can be different. For online layers, the provider of
the imagery layer controls many aspects of the imagery layer, which affects the user's ability to
download, visualize, or use it in analysis. ArcGIS Living Atlas of the World imagery layers are
provided by Esri for users to support their studies and form the foundation for shared
knowledge about the earth. ArcGIS Online also provides additional information that other
organizations have shared to inform and further additional studies. Your organization's data
combined with some of these additional layers can create new insights and allow for better-
informed decision-making. Through ArcGIS Image for ArcGIS Online, your organization can
provide imagery layers that are specific for your organization and can be used in imagery
analysis.
For more information about ArcGIS Image for ArcGIS Online, see the ArcGIS Image for
ArcGIS Online product page (https://links.esri.com/ArcGISImageOnline | www.esri.com/en-us/
arcgis/products/arcgis-online-imagery/overview).

Exercise scenario
This exercise is a continuation of the previous exercise in Section 2. After reviewing your
vegetation mask based on Landsat imagery, you have decided to refine the vegetation layer
with higher-resolution imagery to ensure that the calculations are as good as they can be. The
Sentinel-2 satellite data offers higher-resolution data with similar multispectral characteristics.
A Sentinel-2 raster has been acquired, and your organization has published a multispectral
imagery layer for you to use for your vegetation area analysis. In this exercise, you will use an
ArcGIS Online imagery layer to refine your vegetation mask in ArcGIS Pro to provide more
accurate measurements of vegetation area.

Copyright © 2021 Esri. All rights reserved. 1

 Imagery in Action

Step 1: Add an imagery layer to the map

Now that you have decided to use Sentinel imagery data in your analysis, you have been
given a new imagery layer to use.

a If necessary, start ArcGIS Pro and open the WaterImpactPolicy project.

b In the Catalog pane, click the Project tab, expand the Maps section, and then right-click
the current map and choose Rename.

c Type Preliminary Study and press Enter.

d Close the map, but keep the project open.

In ArcGIS Pro, you can add multiple maps to a single project. Because you want to capture
the progress made, starting a new map will freeze the previous analysis and allow you to start
with the new data.

e At the top of the Catalog pane, click the Portal tab.

f Click the My Organization button , and then in the search field, type sentinel and press
Enter.

g Right-click the Paradise Valley Multispectral imagery layer , point to Add To New, and
then choose Map.

Copyright © 2021 Esri. All rights reserved. 2

 Imagery in Action

The Paradise Valley Multispectral imagery layer is based on a Sentinel-2 raster dataset. When
the imagery layer was created in ArcGIS Image for ArcGIS Online, the multispectral 10
template was used to create the imagery layer, which controls which bands are added to the
imagery layer and the default visualization. Each configuration option during the publishing
affects the bands that are used for the imagery layer, which can affect the spatial resolution of
the imagery layer.
Note: ArcGIS Image for ArcGIS Online will be covered in more detail later in the MOOC. The
configuration options are similar to the options available for mosaic datasets in ArcGIS Pro.

h In the Contents pane, open the Paradise Valley Multispectral imagery layer properties.

i Click the Source tab and expand Raster Information.

1. How many bands are present in the imagery layer?

_______________________________________________________________________________

2. What is the cell size of the imagery layer?

_______________________________________________________________________________

j Click OK to close the layer properties.

Copyright © 2021 Esri. All rights reserved. 3

 Imagery in Action

k Review the following table on Sentinel-2 raster bands.

Sentinel-2 raster bands

Band Description Spatial resolution (meters)

1 Coastal Aerosol 60
2 Blue 10
3 Green 10
4 Red 10
5 Vegetation Red Edge 20
6 Vegetation Red Edge 20
7 Vegetation Red Edge 20
8 Near Infrared 10
8a Narrow Near Infrared 20
9 Water Vapor 60
10 Shortwave Infrared - Cirrus 60
11 Shortwave Infrared - 1 20
12 Shortwave Infrared - 2 20

When choosing a band combination for your analysis, consider the spatial resolution of the
raster bands based on the desired feature. For vegetation analysis, you will use the Near
Infrared and Red raster bands, which have a spatial resolution of 10 meters. So, the band
combination for the imagery layer that you will use is considered to be finer and have a higher
spatial resolution than the Landsat imagery layer used before.
The dynamic Sentinel imagery that you have added to the map has been created for the
vegetation analysis using only the high-resolution bands (Red, Blue, Green, Near Infrared). To
maintain the higher spatial resolution, only the raster bands that have the higher spatial
resolution have been added. If another study requires the shortwave-infrared layers, then a
new dynamic imagery layer will need to be created with those bands. However, this imagery
layer will have a spatial resolution of 20 meters to reflect the addition of the new bands.

Copyright © 2021 Esri. All rights reserved. 4

 Imagery in Action

Step 2: Add contextual layers

In this step, you will add the contextual layers that you created or worked with previously in
the MOOC that will focus the study in Paradise Valley, Arizona, that you researched before.

a In the Catalog pane, click the ArcGIS Online button , and then in the Search ArcGIS
Online field, type potential water owner:esritrainingsvc and press Enter.

b Right-click the Potential Area For Water Policy feature layer and choose Add To
Current Map.

c Click the My Content button , and then locate and add your NDVI Mask For Water
Policy imagery layer to the map.

Note: This exercise uses the imagery layer that was created in a previous exercise. If you did
not complete the Section 1 exercise, then you can add the NDVI Mask For Water Policy
Solution layer provided by clicking the ArcGIS Online button and searching for ndvi mask
owner:esritrainingsvc.
The additional layers overlay the Sentinel imagery layer.

d In the Contents pane, right-click WaterPolicyStudyArea and choose Zoom To Layer.

Copyright © 2021 Esri. All rights reserved. 5

 Imagery in Action

For this imagery layer, you can modify the band combination in the Symbology pane or with
the Band Combination button in the Image Service Layer options.

e In the Contents pane, open the Paradise Valley Multispectral layer properties.

f From the Processing Templates tab, click the Processing Template down arrow to see the
options.
3. What are the processing templates listed for the imagery layer?

_______________________________________________________________________________
The imagery layer from ArcGIS Online is a different type of layer than the layer provided by
ArcGIS Living Atlas. Due to the size and scale of coverage for the Sentinel-2 Views layer, the
imagery layer cannot be manipulated as a normal raster dataset added to ArcGIS Pro would.
However, the imagery layer from ArcGIS Image for ArcGIS Online is a different type of layer
and can be manipulated like a normal raster dataset. When the imagery layers are created
with ArcGIS Image for ArcGIS Online, they can be used for geoprocessing operations and
raster functions.

g Click Cancel to close the layer properties.

h In the Contents pane, turn off the visibility of the NDVI Mask layer.

i Select the Paradise Valley Multispectral imagery layer, and then click the Appearance tab.

j In the Rendering group, click Band Combination and choose Color Infrared.

Copyright © 2021 Esri. All rights reserved. 6

 Imagery in Action

The imagery layer was created with the Sentinel-2 raster type, allowing ArcGIS Pro to use the
metadata for the imagery layer to create band combinations based on the band metadata.
The visualization of the imagery layer uses the Color Infrared band combination but is very
bright. To fix the brightness, you will use the Dynamic Range Adjustment button.
Dynamic range adjustment will use only statistics calculated from the pixels displayed on your
screen for the display. For darker visualizations, this will brighten the display, and for brighter
visualizations, it will reduce the brightness.

k In the Contents pane, ensure that the Paradise Valley Multispectral layer is selected.

l On the Appearance tab, in the Rendering group, click DRA.

Copyright © 2021 Esri. All rights reserved. 7

 Imagery in Action

The color-infrared visualization with DRA allows vegetation to be more visible, and the
contrast within the imagery layer is more vivid.

Step 3: Process rasters with functions

Now that you have added the dynamic imagery layer, you are ready to use raster functions to
analyze the study area and discover the refined vegetation area.
In ArcGIS Pro, you have many methods to analyze your imagery. Geoprocessing tools can be
used to create new output based on your particular analysis. You can also use raster functions
to perform similar analyses but not create output each time. Raster functions are on-the-fly
processing with instructions on how to display the raster in the map and can be combined to
create complex visualizations. When raster functions are used in a map, the output layer is
added to the Contents pane and can be treated as an individual layer. There are benefits to
each processing method, and both methods can be used separately or together to create
your desired output.
For more information about raster functions, see ArcGIS Pro Help: Raster functions.

Copyright © 2021 Esri. All rights reserved. 8

 Imagery in Action

Processing Benefits
method
Geoprocessing
tools • Creates new data on disk
• Easily view and edit geoprocessing history
• Create and use custom geoprocessing tools
• Use batch geoprocessing
• String together processes for complex modeling
• Save sets of geoprocessing tools in your project
• Can be run in the Enterprise environment to use distributed
processing

Raster functions
• No new dataset created
• Fast and efficient; on the fly
• Easily view and edit raster function history
• Create and use custom raster functions
• Chain together processes for complex modeling
• Save sets of raster functions in your project
• Generate processing templates for image services
• Can be run in the Enterprise environment to use distributed
processing

a In the Contents pane, ensure that the Paradise Valley Multispectral imagery layer is still
selected.

b On the Imagery tab, in the Tools group, click Indices and choose NDVI.

c In the NDVI window, specify the following parameters:

• Near Infrared Band Index: 4 - B8

• Red Band Index: 3 - B4

Copyright © 2021 Esri. All rights reserved. 9

 Imagery in Action

The raster bands referred to in the NDVI window are based on the metadata for the imagery
layer. For this imagery layer, there are four bands in the imagery layer based on the Sentinel-2
raster, which had 11 bands.

d Click OK.

Note: It may take a moment for your map to update.

4. In the Contents pane, what are the values reported in the NDVI_Paradise Valley
Multispectral layer?

_______________________________________________________________________________
The values reported in the layer indicate the measure of the Vegetation Index similar to the
calculation in the Landsat Viewer app.

e Right-click the NDVI_Paradise layer and choose Edit Function Chain.

5. In the raster function template editor, which raster function is used to create the
visualization?

_______________________________________________________________________________

f Double-click the Band Arithmetic function to see the calculation used.

The Method is set to NDVI and the Band Indexes are mentioned as the other parameter. The
output calculation is also indicated below the Band Indexes.

Copyright © 2021 Esri. All rights reserved. 10

 Imagery in Action

6. What is the output calculation indicated?

_______________________________________________________________________________

Note: The output specifies the calculation used in the raster function.
This information indicates that both the Near Infrared and Red bands are used in the
calculation as defined by the NDVI method. The Band Arithmetic function is capable of using
known calculations, such as NDVI and custom calculations input by users.
For more information about the Band Arithmetic function, see ArcGIS Pro Help: Band
Arithmetic function.

g Click Cancel to close the Band Arithmetic Properties dialog box, and then close the raster
function template editor.

Step 4: Focus analysis results

In this step, you will focus the vegetation results on the study area with raster functions.

a On the Imagery tab, in the Analysis group, click Raster Functions to open the Raster
Functions pane.

b In the Raster Functions pane, search for clip.

c Click the Clip function to open the properties of the function.

d In the Clip Properties, specify the following parameters:

• Raster: NDVI_Paradise Valley Multispectral

• Clipping Geometry/Raster: WaterPolicyStudyArea
• Check the Use Input Features For Clipping Geometry box.

e Leave the remaining defaults and click Create New Layer.

Copyright © 2021 Esri. All rights reserved. 11

 Imagery in Action

f Turn off the visibility of the full NDVI_Paradise Valley layer.

The raster layer has been clipped to the extent of the study area.
7. In the Contents pane, what are the values reported for the clipped NDVI layer?

_______________________________________________________________________________

The values reported in the Contents pane are based on the values in the layer. For the larger
NDVI layer, those values are in the visible area in the map. Now that the spatial extent of the

Copyright © 2021 Esri. All rights reserved. 12

 Imagery in Action

layer has changed, the values in the Contents pane will change. The new raster layer based
on the study area contains a different subset of values when compared to the entire dataset.

Step 5: Create a new mask

In this step, you will create the refined vegetation mask using similar settings from the
previous study. For the vegetation mask created with the Landsat Viewer app, a vegetation
index value of 0.45 was considered relevant for indicating which areas are considered to
contain vegetation.
Now that you have focused the NDVI results on the study area, you are ready to create the
vegetation mask.

a In the Raster Functions pane, search for greater than.

b Click the Greater Than function, and then in the Greater Than Properties, specify the
following parameters:

• Raster: Clip_NDVI_Paradise Valley Multispectral

• Raster2: 0.45

You are specifying a value that will be used to generate a constant raster where each cell
holds that value. The Greater Than raster function will compare the clipped NDVI raster to
that value and in the output, it will assign a value of 1 to each cell with a value greater than
0.45 and a value of 0 to each cell with a value less than 0.45.
Note: The value of 0.45 was previously established as the threshold value to evaluate the
NDVI results in the Landsat Explorer app.

c Leave the remaining defaults and click Create New Layer.

Copyright © 2021 Esri. All rights reserved. 13

 Imagery in Action

The output is visualized with the Stretch renderer. The values in the output are either 0 or 1 to
indicate which areas are above the threshold of 0.45. To create a mask that compares to the
previous vegetation mask, you will change the symbology of the raster layer.

d In the Contents pane, select the Greater Than_Clip_NDVI layer, if necessary.

e On the Appearance tab, in the Rendering group, click Symbology to open the Symbology
pane.

f In the Symbology pane, change the Primary Symbology from Stretch to Classify.

g When prompted to create a histogram, click Yes.

Copyright © 2021 Esri. All rights reserved. 14

 Imagery in Action

The classified renderer will list two classes. One class will have a value of 1 and the other a
value of 0. These classes indicate values that exceed the threshold and values that do not.

h In the lower section of the Symbology pane, on the Classes tab, click the color swatch for
the 0 value and choose No Color.

i Update the color swatch for the 1 value to a green color.

The vegetation mask layer will display as green and the areas that are not part of the mask will
not be displayed.

Copyright © 2021 Esri. All rights reserved. 15

 Imagery in Action

Step 6: Review area calculation

In this step, you will compare the vegetation mask that you created with Sentinel imagery with
the mask created with Landsat imagery.

a In the Contents pane, turn off the visibility of the Clip_NDVI_Paradise Valley layer and turn
on the visibility of your NDVI Mask For Water Policy layer.

You can see the blue of the NDVI mask around some of the edges of the new mask.

b Zoom in to the large area in the middle.

Copyright © 2021 Esri. All rights reserved. 16

 Imagery in Action

Due to the differences in the spatial resolution between the Landsat imagery and the Sentinel
imagery, the new vegetation mask appears to cover a smaller area than the previous mask. To
refine the estimate of the area, you will need to survey the total area covered by the new
mask. To see the total area covered, you will need to generate a raster attribute table to see
the total area covered by the Greater Than raster layer. The output of the raster function does
not have a raster attribute table, so you will export the mask to create a raster attribute table.

c In the Contents pane, select the Greater Than_Clip_NDVI layer, if necessary.

d On the Data tab, in the Tools group, click Export Raster.

e In the Export Raster pane, for Output Raster Dataset, click the Browse button .

f Browse to C:\EsriMOOC\Data, and then name the output NDVI_RefinedMask.tif and

click Save.

g Accept the remaining defaults and click Export.

h From the Catalog pane, click the Project tab and then under Folders, right-click Data and
choose Refresh to see the NDVI_RefinedMask.tif file.

i In the Contents pane, right-click NDVI_RefinedMask.tif and choose Zoom To Layer.

Copyright © 2021 Esri. All rights reserved. 17

 Imagery in Action

Now that you have created the new raster dataset, you are ready to create the raster attribute
table to measure the area. Before processing, it is good practice to remove the raster to
prevent the table from locking and not updating.

j Remove the NDVI_RefinedMask.tif layer from the map.

Hint: Right-click > Remove
After you build the raster attribute table, this layer will be added to the Contents pane again.

k On the Analysis tab, in the Geoprocessing group, click Tools to open the
Geoprocessing pane.

l In the Geoprocessing pane, search for build raster and choose Build Raster Attribute
Table.

m In the Build Raster Attribute Table tool, for the Input Raster, choose
NDVI_RefinedMask.tif.
Note: If you do not see this file in the drop-down list, click the Browse button , browse to
C:\EsriMOOC\Data, select NDVI_RefinedMask.tif, and click OK.

n Click Run.

Copyright © 2021 Esri. All rights reserved. 18

 Imagery in Action

Note: The symbology colors in your map may differ from the preceding graphic.
The raster layer symbology changes due to the presence of the raster attribute table. The
raster layer will be visualized using the Unique Values renderer.

o In the Symbology pane, change the colors of the values as you did before, making the 0
value display as No Color and the 1 value display as green again.

p In the Contents pane, turn off the visibility of the Greater Than_Clip_NDVI layer.

q Right-click the NDVI_RefinedMask.tif layer and choose Attribute Table.

Note: You can also press Ctrl and double-click a layer to see its attribute table.
8. What is the reported count for the Value 1?

_______________________________________________________________________________

The value reported in the raster attribute table indicates the number of cells that have a value
of 1. To calculate the area, the cell size multiplied by the count indicates the total area.
Note: Remember that the cell size is 10 meters, which indicates that the area covered by one
cell is 10 m x 10 m or 100 m.
9. What is the refined calculated vegetation area based on the raster attribute table?

_______________________________________________________________________________

_______________________________________________________________________________

Copyright © 2021 Esri. All rights reserved. 19

 Imagery in Action

In the Landsat-based area calculation, the vegetation mask estimated an area of 4.8 sq.
kilometers for comparison.

r Close the attribute table.

s In the Contents pane, turn off the visibility of the Paradise Valley Multispectral imagery
layer and drag your NDVI Mask For Water Policy layer above the new
NDVI_RefinedMask.tif layer.

In the map, you can see the additional areas visualized in green polygons in the new
vegetation mask. These new areas found explain how the estimated vegetation area has
changed.
Because the new vegetation mask is based on higher-resolution imagery, you were able to
identify smaller vegetation areas not included in the Landsat vegetation mask. When
presenting this data to your manager, you now have a good estimate of the overall total
vegetation area and may decide to further refine the vegetation area calculation to remove
some of the new smaller areas discovered in favor of larger areas. Either way, you have
created information for decision makers to consider when making any water policy changes.

t Save the project and exit ArcGIS Pro.

Copyright © 2021 Esri. All rights reserved. 20

 Imagery in Action

Answers to Exercise Questions

1. How many bands are present in the imagery layer?

There are four bands present in the imagery layer.

2. What is the cell size of the imagery layer?

The listed cell size is 10 meters.

3. What are the processing templates listed for the imagery layer?
There are no processing templates listed in the drop-down list.

4. In the Contents pane, what are the values reported in the NDVI_Paradise Valley
Multispectral layer?
The values are 1 to -0.4229.

5. In the raster function template editor, which raster function is used to create the
visualization?
The Band Arithmetic function is used to calculate the NDVI.

6. What is the output calculation indicated?

(NIR - Red) / (NIR + Red) is the calculation.

7. In the Contents pane, what are the values reported for the clipped NDVI layer?
The values reported now are 0.747 to -0.235.

8. What is the reported count for the Value 1?

The value is around 75,000.

9. What is the refined calculated vegetation area based on the raster attribute table?
75,000 x 100 sq. meters (cell size 10 x 10) = 7,500,000 sq. meters or 7.5 sq
kilometers. (Note that your calculation will vary slightly from this example.)

Copyright © 2021 Esri. All rights reserved. 21