GEOG 205 – Lab 08 | Satellite / Google Earth

1. Examining Satellite Images

In this section, you will be working with Sentinel-2 satellite imagery. Where most images you see contain Blue, Green and Red (more commonly known as RGB), the Sentinel-2 images contain 13 bands. The full list of bands is available from the European Space Agency: https://sentinel.esa.int/web/sentinel/user-guides/sentinel-2-msi/resolutions/radiometric

There is also this handy reference table which provides names to the bands from Hatari Labs: https://www.hatarilabs.com/ih-en/how-many-spectral-bands-have-the-sentinel-2-images

As humans we only see 3 colours, Blue, Green, and Red, however, we can use the computer to move other bands into one of these spectra that we can see. Let’s take a look at a satellite image. If you want to acquire your own satellite images they can be downloaded free of charge from Open Access Hub: https://scihub.copernicus.eu/dhus/#/home

When you download the imagery it actually comes as a collection of files. In windows file explorer, take a look at this directory:

L:\GEOG205\lab08\MtRobson\S2B_MSIL2A_20201002T191229_N0214_R056_T11ULU_20201002T215256.SAFE

If you look around inside of the GRANULE subfolder, you will find the images sorted into folders for each resolution, and a JPEG 2000 (.jp2) file for each band. However we do not want to add bands one file at a time, we will make use of meta files, files that store information about the other files.

If you are really nerdy and what to know what a meta file looks like you can open MTD_MSIL2A.xml in notepad.

For the metafiles to work correctly we must add the file through the catalogue pane (dragging and dropping from windows file explorer will not produce the desired results!).

Start a new project and add a new folder connection for L:\GEOG205\lab08

As you expand the directory tree in ArcGIS Pro you will notice a layer of files below the XML files, these represent the processes for opening the imagery.

These all represent different views of the data useful for a variety of Remote Sensing tasks (Take GEOG 357 if you want to learn more).

For now we are only interested in the BOA (Bottom of Atmosphere) Reflectance, there are 4 variations of this, the first one contains all 13 bands, and the following options contain only the bands of the specified resolution.

Add the BOA Reflectance that contains all 13 bands to your project. (This is a relatively large set of files, just under 1GB, so give it a few seconds to load). If all goes according to plan you should see the below.

Next, we will look at some ways of altering the appearance of this image. With the image selected open the ‘Raster Layer’ tab. We are interested in the Rendering Section (*If your image opened very dark set the gamma to 2.2 in the enhancement).

• 
  * All the data for this lab has already been prepared for stretching, a process that took roughly 45 minutes to complete (a press start and work on something else sort of process), if you were to download a fresh set of imagery, and it appears very dark (or totally black), you would need to right click the layer and “Calculate Statistics”

• Stretch Type:
  In a typical image you see on the computer screen the image is a grid of pixels, each having a number representing each colour, and these numbers are generally 8bit (some high end monitors support 10bit) or range from 0-255, representing the brightness of that colour. The data from Sentinel is a little different in that it is 12bit (2^12) data ranging from 0-4095, the stretch affects how we map these values down to the 0-1023 range our new monitors can display or the 0-255 range that most monitors can display.

Setting the stretch to none performs a linear mapping such that 4095 -> 255, 2048 -> 128, and 0 -> 0. But what if the data in our image doesn’t actually extend across the 0-4095 range?

This is where stretching comes in, it allows us to make the most of the displays range for the example given:

• DRA:
  Dynamic Range Adjustment: when this option is enabled, the stretch is performed over only the pixels on the screen. This allows contrast to increase as you zoom in on the image.
• Resampling Type:
  When zoomed out on the map, there will be multiple pixels mapped to each pixel on the screen. The resampling type affects how they are mapped. The simplest form is the nearest neighbour. In this strategy, the pixel that is closest to the center of the pixel on the display is shown. The other strategies will provide averages of the pixels in the neighbourhood, producing smoother and potentially more accurate results.
• Band Combination:
  Band Combination provides quick access to commonly used arrangement bands, listed by use. Because of the metafiles, ArcGIS Pro understands what each band represents (this will change with various satellite platforms), and this provides a quick way of getting a relatively consistent display. Try looking at a few and see how the different features pop.

• Brightness / Contrast / Gamma:
  These provide some basic controls in how the image is displayed on the monitor. Note here that brightness and contrast have neutral values of 0 and 10 respectively, where the gamma default is at 2.2 (There is no need to understand why, but if you are curious there is more information on gamma available here: https://en.wikipedia.org/wiki/Gamma_correction).

Exporting data for use in other software

Zoom into Mt Robson in the North centre of the map (see below image for a reference of the area that should be visible.)

Now we are going to export as Raster (Right click on the layer in drawing order, go to Data and Export Raster)

Export Raster

1. Set the output Raster Dataset to a location inside your lab 8 folder. And name the file with a .jpg file extension.
2. Set the desired coordinate reference system: WGS 1984 (4326).
3. In the Clipping Geometry drop down select “Current Display Extent”
4. Select ‘Use Renderer’: This option will preserve the stretch and band combination we have chosen.
5. Force RGB, makes sure only 3 bands are exported.

2. Working with Google Earth

Google Earth overlays satellite imagery, aerial photography and GIS information over a 3D model of the Earth. It allows users to search for addresses, enter coordinates, or simply use the mouse to browse to a location.  Google Earth enables the user to view satellite images, layers of data provided by Google and by the Internet community at large. Google Earth also has 3D terrain data collected by NASA’s space shuttle and available in the public domain. In addition, Google has provided a layer allowing one to see 3D buildings for some major cities.

On your personal computer, the Google Earth application can be downloaded from https://www.google.com/earth/versions/#earth-pro, but the lab computers already have Google Earth installed.

Add Imagery to Google Earth Pro

Open Google Earth Pro (desktop icon, or from the start menu).

Drag the image L:\GEOG205\lab08\MtRobson\robson_1.jpg from Windows File Explorer to Google Earth, you will see a message saying flying to location.

You will then be prompted with a message saying the image is larger than the maximum supported size, choose to Create Super Overlay, and in the file browser that opens create a new folder called robson_overlay inside of your lab08 folder and select that as the location for the overlay. (If the super overlay takes too long, you can cancel and attempt again with Scale, this will not be as sharp but will process faster).

Once it has had time to make the super overlay you should see it appear inside of Google Earth.

You should notice that 3D navigation is substantially faster than it was in ArcGIS Pro, however you may also notice some artifacting in the images. So some pros and some cons.

Annotations with Google Earth

Google Earth is most commonly used for quickly marking up imagery. Above the map you will see tools for creating Placemarks (points), Polygons, and Paths.

Start by clicking on the yellow push pin, and then click somewhere on the map, you will see the pin placed, and a dialogue opens.

1. The name field is used to label the point you just placed.
2. You are able to change the symbol on the map from a gallery of symbols.
3. Description is used to add additional information about the placemark that can be seen by clicking on the point.
4. Style and Color provide some additional options related to the symbol and label.

One more tip: If the placemark is not quite where you want it, you can click and drag it over the map.

A note on polygon features, due to the way the layers are rendered, the polygons will always be drawn below the draped imagery. For this reason if you are using draped imagery (like the 2.5D look we are using in Google Earth now), polygons should be avoided.

Historical Imagery

Another very helpful feature in google earth is the ability to view historical scenes, this provides an easy way to look at changes over time on the landscape, this can be done with the clock button.

Take a look at your favorite vacation spot, and see how the landscape has changed over the years.

Adding other layers to the map:

In the lower left corner you will see a layers panel, you can use this to add labels and placemarkers for a variety of different data sources.

Exporting Your Layers

You can export individual features by right clicking and choosing Save Place As…

You will be given options for KML or KMZ (KMZ is a compressed KML)

You can also select a folder and Save the Entire Folder as a KML allowing for multiple features or layers to be exported to a single file.

Exporting your image map

Exporting maps from Google Earth is very simple compared to ArcGIS Pro, but also comes at the cost of customization.

1. Clicking on the print icon will open the rest of your print options.
2. Page Setup is where you pick your page size, I have found this to always be a good place to start, before trying to get the perfect view.
3. Add a title, as well as a description including your name and date the map was produced.
4. You can turn off any features from the Legend should you not want them.
5. Either send it to your printer or save it as a PDF.

Assignment 8:

This week’s assignment is split into two parts. Both parts should be submitted as a PDF file using landscape oriented letter sized paper.

Part 1:

In the assignment folder, there is a Sentinel-2 Image from West of Edmonton Alberta. For this assignment, you will need to make a simple map using only this image as a single layer. However, the Legend will be less straightforward than normal, as ArcGIS will not create a legend for you. And you will need to utilize the Graphics and Text features to manually create a legend that will provide meaning to the readers of your map.

Zoom in on a portion of the map where the fields can be clearly seen (roughly 1:100,000), you can pick any area that looks interesting. And set the stretch and band combination to show the Agriculture well.

HINT: if the colours on your layout look darker than your map, try “Esri” for Stretch type.

Using the agricultural band combination, water will appear as black, vegetation green, and bare earth as brown (possibly with a magenta tint). Use the Graphics and Text controls to make a legend to show this. One possible solution is to make three rectangles filling each with a colour similar to what appears on the map, and putting a text label beside.

Make the Layout for Landscape Letter size paper; including, your map frame, a handmade legend, Title, Name, Date, Scale Bar, and Locator Map (You may wish to use the ‘province boundary’ layer in the UNBC ArcGIS Portal).

Export your map as a PDF.

Part 2:

Create an image map in Google Earth that shows the retreat of the the Swift Current Glacier.

Inside of the assignment folder you will find robson-glacier-extents. You will need to convert these shapefiles to KML/KMZ before opening in Google Earth. (You will find Layer to KML in the Analysis Toolbox https://pro.arcgis.com/en/pro-app/latest/tool-reference/conversion/layer-to-kml.htm)
*Note that in Google Earth, the Drawing order is not determined by the legend, rather the order in which the files are added to Google Earth, with the most recent file on top.

Your image map should include the usual: Title, Description, Name, Date, and Legend.

Upload the two PDFs to Moodle.