update

docs/qgis-cloud-native-geospatial.html

@@ -319,6 +319,12 @@ <h4 class="author">Ujaval Gandhi</h4>
 id="toc-hosting-cloud-native-geospatial-data">4.2 Hosting Cloud Native
 Geospatial Data</a></li>
 </ul></li>
+<li><a href="#supplement" id="toc-supplement">Supplement</a>
+<ul>
+<li><a href="#analyzing-cloud-hosted-satellite-imagery"
+id="toc-analyzing-cloud-hosted-satellite-imagery">Analyzing Cloud-Hosted
+Satellite Imagery</a></li>
+</ul></li>
 <li><a href="#data-credits" id="toc-data-credits">Data Credits</a></li>
 <li><a href="#license" id="toc-license">License</a></li>
 </ul>
@@ -331,7 +337,7 @@ <h4 class="author">Ujaval Gandhi</h4>
 <div style="page-break-after: always;"></div>
 <div id="introduction" class="section level1">
 <h1>Introduction</h1>
-<p>This workshop will introduce participants to the modern approach to
+<p>This workshop introduces participants to the modern approach to
 working with large datasets in QGIS. Modern data formats - such as
 Cloud-Optimized GeoTIFFs (COG), Cloud-Optimized Point Clouds (COPC), and
 FlatGeoBuf (FGB) allow datasets to be streamed from cloud storage
@@ -774,7 +780,7 @@ <h2>3.2 Analyze Landcover Change</h2>
 <p><img src="images/qgis_cloud_native_geospatial/landcover14.png" width="75%" style="display: block; margin: auto;" /></p>
 <ol start="15" style="list-style-type: decimal">
 <li>The <code>waterlost_export</code> is a local raster with pixels
-showing regions that experienced loss in surface water. Let’s conver
+showing regions that experienced loss in surface water. Let’s convert
 this layer to polygons. From the Processing Toolbox, search and locate
 the <strong>GDAL → Raster Conversion → Polygonize (raster to
 vector)</strong> algorithm.</li>
@@ -812,13 +818,211 @@ <h2>4.1 Creating Cloud Native Geospatial Data</h2>
 <h2>4.2 Hosting Cloud Native Geospatial Data</h2>
 </div>
 </div>
+<div id="supplement" class="section level1">
+<h1>Supplement</h1>
+<div id="analyzing-cloud-hosted-satellite-imagery"
+class="section level2">
+<h2>Analyzing Cloud-Hosted Satellite Imagery</h2>
+<p>In this section, we will learn how to query a STAC catalog of
+Sentinel-2 L2A data, load and create a RGB composite and compute a
+spectral index using a cloud-native approach.</p>
+<ol style="list-style-type: decimal">
+<li>Start by loading a basemap. From the <em>Browser</em> panel, scroll
+down and locate <strong>XYZ Tiles → OpenStreetMap</strong> tile layer.
+Drag and drop it to the main canvas. Zoom to your region of interest. We
+will use the <em>STAC API Browser</em> plugin to query for satellite
+images over this area. Open the plugin from <strong>Plugins → STAC API
+Browser Plugin → Open STAC API Browser</strong>.</li>
+</ol>
+<p><img src="images/qgis_cloud_native_geospatial/satellite1.png" width="75%" style="display: block; margin: auto;" /></p>
+<ol start="2" style="list-style-type: decimal">
+<li>Select <code>Earth Search</code> from the <em>Connections</em>
+dropdown. Click <em>Edit</em> to view the connection details.</li>
+</ol>
+<p><img src="images/qgis_cloud_native_geospatial/satellite2.png" width="75%" style="display: block; margin: auto;" /></p>
+<ol start="3" style="list-style-type: decimal">
+<li>Make sure the URL is
+<code>https://earth-search.aws.element84.com/v1</code> and click
+<em>OK</em>.</li>
+</ol>
+<p><img src="images/qgis_cloud_native_geospatial/satellite3.png" width="75%" style="display: block; margin: auto;" /></p>
+<ol start="4" style="list-style-type: decimal">
+<li>In the <em>STAC API Browser</em> window, click the <em>Fetch
+collections</em> button. Once the list updates, select the
+<code>Sentinel-2 Level-2A</code> collection. If you do not see this
+dataset listed, please go to the previous step and verify the URL is
+correct.</li>
+</ol>
+<p><img src="images/qgis_cloud_native_geospatial/satellite4.png" width="75%" style="display: block; margin: auto;" /></p>
+<ol start="5" style="list-style-type: decimal">
+<li>Check the <em>Filter by date</em> option and enter a start and end
+date. Sentinel-2 Level-2A dataset is available from 2017 onward and a
+new image is collected every 5 days. Check the <em>Extent</em> option
+and click <em>Map Canvas Extent</em> button to fill the coordinates of
+the region visible on the map. Click <em>Search</em>.</li>
+</ol>
+<p><img src="images/qgis_cloud_native_geospatial/satellite5.png" width="75%" style="display: block; margin: auto;" /></p>
+<ol start="6" style="list-style-type: decimal">
+<li>In the <em>Results</em> tab, you will see the list of matching
+images. From the preview of the image, locate a reasonably cloud-free
+image and click <em>View assets</em>.</li>
+</ol>
+<p><img src="images/qgis_cloud_native_geospatial/satellite6.png" width="75%" style="display: block; margin: auto;" /></p>
+<ol start="7" style="list-style-type: decimal">
+<li>For this exercise, we will load data for 4 image bands. Check the
+<em>Select to add as a layer</em> box for <code>Blue (band 2)</code>,
+<code>Green (band 3)</code>, <code>Red (band 4)</code> and
+<code>SWIR 1 (band 11)</code>. Click <em>Add select assets as layers
+(4)</em>.</li>
+</ol>
+<p><img src="images/qgis_cloud_native_geospatial/satellite7.png" width="75%" style="display: block; margin: auto;" /></p>
+<ol start="8" style="list-style-type: decimal">
+<li>The selected bands will be loaded as new layers in the main QGIS
+canvas. Each band is a Cloud-Optimized GeoTIFF (COG) that is hosted on
+AWS and being streamed to your QGIS. We can combine multiple bands into
+a single image and visualize it. These are called <em>Color
+Composites</em>. Let’s stack the Red, Green and Blue bands to create an
+natural color composite. Go to <strong>Processing → Toolbox</strong>.
+From the <em>Processing Toolbox</em>, search and locate the <strong>GDAL
+→ Raster miscellaneous → Build virtual raster</strong> algorithm.
+Double-click to open it.</li>
+</ol>
+<p><img src="images/qgis_cloud_native_geospatial/satellite8.png" width="75%" style="display: block; margin: auto;" /></p>
+<ol start="9" style="list-style-type: decimal">
+<li>This algorithm creates a new layer in the <a
+href="https://gdal.org/en/latest/drivers/raster/vrt.html">Virtual
+Raster</a> format which allows you to combine multiple raster layers in
+to a single file without consuming extra disk space. Select
+<code>Blue (band 2)</code>, <code>Green (band 3)</code>,
+<code>Red (band 4)</code> as <em>Input layers</em> and check <em>Place
+each input file into a separate band</em>. Click the <code>...</code>
+button next to <em>Virtual</em> and save the output file as
+<code>rgb.vrt</code>. Next click <em>Run</em>.</li>
+</ol>
+<p><img src="images/qgis_cloud_native_geospatial/satellite9.png" width="75%" style="display: block; margin: auto;" /></p>
+<ol start="10" style="list-style-type: decimal">
+<li>A new layer <code>rgb</code> will be added to the <em>Layers</em>
+panel. This layer contains references to the 3 different images. Note
+that the order of the bands in alphabetical, so the mapping in the
+virtual raster is as follows: <em>Band 1 → Blue (band 2)</em>, <em>Band
+2 → Green (band 3)</em> and <em>Band 3 → Red (band 4)</em>. Open the
+<em>Layer Styling Panel</em> and place <code>Band 3</code>,
+<code>Band 2</code> and <code>Band 1</code> as <em>Red</em>,
+<em>Green</em> and <em>Blue</em> bands. Change the <em>Min</em> and
+<em>Max</em> values to <code>0</code> and <code>3000</code>
+respectively. You will see the image now appear in natural colors.</li>
+</ol>
+<p><img src="images/qgis_cloud_native_geospatial/satellite10.png" width="75%" style="display: block; margin: auto;" /></p>
+<ol start="11" style="list-style-type: decimal">
+<li>We can also do some analysis with the satellite image bands. Let’s
+calculate the <strong>Modified Normalized Difference Water Index
+(MNDWI)</strong>. This is a widely used spectral index for identifying
+and extracting water pixels from satellite images. This index is
+calculated by taking the normalized ratio of the Green (band 3) and the
+SWIR-1 (band 11) bands. From the <em>Processing Toolbox</em>, search and
+locate the <strong>GDAL → Raster analysis → Raster calculator
+(virtual)</strong> algorithm. Double-click to open it.</li>
+</ol>
+<p><img src="images/qgis_cloud_native_geospatial/satellite11.png" width="75%" style="display: block; margin: auto;" /></p>
+<ol start="12" style="list-style-type: decimal">
+<li>In the <em>Raster calculator (virtual)</em> dialog, click the
+<code>..</code> button for <em>Input layers</em>.</li>
+</ol>
+<p><img src="images/qgis_cloud_native_geospatial/satellite12.png" width="75%" style="display: block; margin: auto;" /></p>
+<ol start="13" style="list-style-type: decimal">
+<li>Select the <code>Green (band 3)</code> and the
+<code>SWIR 1 (band 11)</code> layers and click <em>OK</em>.</li>
+</ol>
+<p><img src="images/qgis_cloud_native_geospatial/satellite13.png" width="75%" style="display: block; margin: auto;" /></p>
+<ol start="14" style="list-style-type: decimal">
+<li>Next click the <em>Expression</em> button.</li>
+</ol>
+<p><img src="images/qgis_cloud_native_geospatial/satellite14.png" width="75%" style="display: block; margin: auto;" /></p>
+<ol start="15" style="list-style-type: decimal">
+<li>Remember the formula for MNDWI is (Green - SWIR1)/(Green + SWIR1).
+You can click on the layer names to enter it in the <em>Raster
+Calculator Expression</em> and construct the expression as shown below.
+Once done, click <em>OK</em>.</li>
+</ol>
+<pre><code>(&quot;Green (band 3) - 10m@1&quot; - &quot;SWIR 1 (band 11) - 10m@1&quot;)
+/ (&quot;Green (band 3) - 10m@1&quot; + &quot;SWIR 1 (band 11) - 10m@1&quot;)</code></pre>
+<p><img src="images/qgis_cloud_native_geospatial/satellite15.png" width="75%" style="display: block; margin: auto;" /></p>
+<ol start="16" style="list-style-type: decimal">
+<li>Enter the <em>Output layer name</em> as <code>MNDWI</code> and click
+<em>Run</em>.</li>
+</ol>
+<p><img src="images/qgis_cloud_native_geospatial/satellite16.png" width="75%" style="display: block; margin: auto;" /></p>
+<ol start="17" style="list-style-type: decimal">
+<li>A new layer <code>MNDWI</code> will be added to the <em>Layers</em>
+panel. Click the open the <em>Layer Styling Panel</em> button. Change
+the renderer to <code>Singleband psuedocolor</code>. Set the
+<em>Min</em> and <em>Max</em> values to <code>0</code> and
+<code>0.8</code> respectively. Choose the <code>Blues</code> as the
+<em>Color ramp</em>. You will see a visualization where all detected
+water pixels are highlighted in the blue color.</li>
+</ol>
+<p><img src="images/qgis_cloud_native_geospatial/satellite17.png" width="75%" style="display: block; margin: auto;" /></p>
+<ol start="18" style="list-style-type: decimal">
+<li>The MNDWI image contains values in the range -1 to +1 with pixel
+values 0 and above typically indicating presence of water. We can
+extract the water pixels to a new layer by applying a threshold of 0.
+From the <em>Processing Toolbox</em>, search and locate the <strong>GDAL
+→ Raster analysis → Raster calculator (virtual)</strong> algorithm.
+Double-click to open it.</li>
+</ol>
+<p><img src="images/qgis_cloud_native_geospatial/satellite18.png" width="75%" style="display: block; margin: auto;" /></p>
+<ol start="19" style="list-style-type: decimal">
+<li>Select the <code>MNDWI</code> layer as the <em>Input layers</em> and
+enter the expression shown below. Set the <em>Output layer name</em> to
+<code>Water</code> and click <em>Run</em>.</li>
+</ol>
+<pre><code>&quot;MNDWI@1&quot; &gt; 0</code></pre>
+<p><img src="images/qgis_cloud_native_geospatial/satellite19.png" width="75%" style="display: block; margin: auto;" /></p>
+<ol start="20" style="list-style-type: decimal">
+<li>A new layer named <code>Water</code> will be added to the
+<em>Layers</em> panel. Click the open the <em>Layer Styling Panel</em>
+button. Set the <em>Min</em> and <em>Max</em> values to <code>0</code>
+and <code>1</code> respectively. You will see all the extracted water
+pixels in white. Remember this image is still a virtual image being
+computed on-the-fly by streaming in the required data directly from the
+cloud-hosted COGs. You can Zoom and Pan the map and the water pixels
+will be extracted dynamically. We can save the results as a local
+GeoTIFF. Right-click on the <code>Water</code> layer and select
+<strong>Export → Save As…</strong>.</li>
+</ol>
+<p><img src="images/qgis_cloud_native_geospatial/satellite20.png" width="75%" style="display: block; margin: auto;" /></p>
+<ol start="21" style="list-style-type: decimal">
+<li>Click the <code>..</code> button next to <em>File name</em> and save
+the file as <code>extracted_water.tif</code>. Expand the <em>Extent</em>
+section and click the <em>Map Canvas Extent</em> button. Set both both
+<em>Horizontal</em> and <em>Vertical</em> values for <em>Resolution</em>
+to be <code>10</code> meters. Check the <em>Create options</em> box and
+select <code>High Compression</code> as the <em>Profile</em>. Click
+<em>OK</em> to start the export process. At this stage, QGIS will fetch
+the required pixels from the cloud dataset at the required resolution,
+perform the computation to calculate MNDWI and extract the water pixels.
+Depending on how big is your region, this process may take a few
+minutes.</li>
+</ol>
+<p><img src="images/qgis_cloud_native_geospatial/satellite21.png" width="75%" style="display: block; margin: auto;" /></p>
+<ol start="22" style="list-style-type: decimal">
+<li>Once complete, a new layer <code>extracted_water</code> will be
+added to the <em>Layers</em> panel. You were able visualize a satellite
+image and perform a complex calculation to extract water from the image
+without first having to download the entire scene. Saving time,
+bandwidth and computation resources.</li>
+</ol>
+<p><img src="images/qgis_cloud_native_geospatial/satellite22.png" width="75%" style="display: block; margin: auto;" /></p>
+</div>
+</div>
 <div id="data-credits" class="section level1">
 <h1>Data Credits</h1>
 <ul>
 <li>Global Edge-matched Subnational Boundaries: Humanitarian Edge
 Matched, FieldMaps, OCHA, geoBoundaries, U.S. Department of State,
 OpenStreetMap. Downloaded from <a href="https://fieldmaps.io/data"
 class="uri">https://fieldmaps.io/data</a></li>
+<li>Sentinel-2 Level-2A: Contains Copernicus Sentinel data.</li>
 </ul>
 </div>
 <div id="license" class="section level1">