Detect ships with SAR imagery

In this tutorial, as an analyst for the Panama Traffic Services agency, you will help determine ship congestion by detecting ships in Synthetic Aperture Radar (SAR) imagery and you’ll use the result to generate a vessel density map. A vessel density map shows the distribution of ships (that is, maritime traffic), based on the number of ships per unit area. SAR satellite sensors send and receive back microwaves to create high-resolution images of the Earth's surface. Unlike traditional methods such as the Automatic Identification System (AIS), which relies on broadcast systems aboard ships, SAR can help detect ships even in areas where AIS data may not be available or reliable. Understanding port congestion by incorporating a SAR-based approach can improve maritime traffic safety and efficiency.

This tutorial was last tested on January 19, 2024, using ArcGIS Pro 3.2. If you're using a different version of ArcGIS Pro, you may encounter different functionality and results.

Requirements

• ArcGIS Pro (see options for software access)
• ArcGIS Image Analyst

Prepare SAR imagery for analysis

First, you'll download and open the project in ArcGIS Pro. You will add a SAR image representing the Panama Canal and its surroundings to the map. Then you will prepare it by applying radiometric calibration to it and by changing its symbology.

Download and open the project

You'll download a project that contains all the data for this tutorial and open it in ArcGIS Pro.

1. Download the Detect_ships_with_SAR_imagery.zip file and locate the downloaded file on your computer.

   Note:

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

2. Right-click the Detect_ships_with_SAR_imagery.zip file and unzip it to a location on your computer, such as drive C.
3. Open the extracted Detect_ships_with_SAR_imagery folder and double-click Detect_ships_with_SAR_imagery.aprx to open the project in ArcGIS Pro.

   

4. If prompted, sign in to your ArcGIS organizational account.

   Note:

   If you don't have access to ArcGIS Pro or an ArcGIS organizational account, see options for software access.

   The project opens.

   

   Currently, the map contains only the default topographic basemap, centered on the Panama Canal. The SAR image you'll use in this tutorial is included in the project; you'll add it to the map.

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

   

   The Catalog pane appears.

6. In the Catalog pane, click the arrow next to Folders to expand it. Similarly, expand the Detect_ships_with_SAR_imagery and Input Data folders.

   

   The Input Data folder contains most of the data you'll use in this tutorial.

7. Expand the SAR Image folder, right-click ICEYE_GRD_SC_244270_20220207T192226.xml, and choose Add To Current Map.

   

   The image appears on the map.

   

   This SAR image represents the Panama Canal and its surroundings and was captured in February 2022.

   Note:

   The image is derived from a SAR Level-1 Ground Range Detected (GRD) data product from the ICEYE satellite imagery provider. Founded in 2014, ICEYE aims to provide reliable and timely information about planet Earth through its constellation of small SAR satellites. These satellites capture high-resolution images of the Earth’s surface, enabling monitoring and analysis of various phenomena, including natural disasters, environmental changes, maritime activity, and infrastructure monitoring. Learn more about ICEYE SAR data. You can also learn more about SAR GRD and other product types commonly used to deliver SAR imagery in SAR Satellite Data.

   You'll give the layer a more meaningful name.

8. In the Contents pane, click the Scan_ICEYE_GRD_SC_244270_20220207T192226 layer once to select it, and click it again to switch to edit mode. Type Panama SAR and press Enter.

   

   Tip:

   Alternatively, to rename a layer, select the layer and press F2 or Fn+F2.

Apply radiometric calibration

You will now prepare the SAR image by calibrating it radiometrically. Radiometric calibration is necessary to obtain meaningful backscatter values that can be related to the physical properties of features in the image. Learn more about the concept of backscatter in the Fundamentals of Synthetic Aperture Radar (SAR) guide. You will perform the calibration with the Apply Radiometric Calibration tool.

Note:

Level-1 SAR data contains uncalibrated digital numbers (DN) as pixel values. The calibration process converts the DN values into radar backscatter intensity values. The backscatter intensity is the proportion of emitted signal striking the surface to the received signal (aka DN) reflected from it. Radiometry or brightness of the backscatter intensity can vary. Therefore, radiometric calibration is important for normalizing these values to obtain accurate qualitative observations and quantitative measurements from the SAR image.

1. On the ribbon, on the Analysis tab, in the Geoprocessing group, click the Tools button.

   

   The Geoprocessing pane appears.

2. In the Geoprocessing pane, click the Toolboxes tab.

   

3. Expand the Image Analyst Tools toolbox to see all the toolsets it contains.

   

4. Under Image Analyst Tools, expand the Synthetic Aperture Radar toolset.

   

   This toolset contains SAR geoprocessing tools, including the Apply Radiometric Calibration tool.

   Tip:

   Alternatively, you can also find the Apply Radiometric Calibration tool by typing its name in the Geoprocessing pane search box.

5. Click the Apply Radiometric Calibration tool to open it.

   

6. Set the following tool parameters:
   • For Input Radar Data, choose Panama SAR.
   • For Output Radar Data, type Panama SAR - Radiometric Calibration.crf
   • For Polarization Bands, confirm the box is checked.
   • For Calibration Type, choose Gamma nought.

   The output will be in the Cloud Raster Format (CRF) file format, which is optimized for writing and reading large raster files in a distributed processing and storage environment.

   You choose the Gamma nought option because the Detect Bright Ocean Objects tool that you’ll use later in the workflow requires a SAR input with that calibration type.

   

   Note:

   Learn more about the calibration options in the Apply Radiometric Calibration tool documentation.

7. Click Run.

   After a few moments, the new Panama SAR - Radiometric Calibration.crf layer appears.

   

Note:

To view more information about the image, in the Contents pane, right-click the Panama SAR - Radiometric Calibration.crf layer, choose Properties, and click the Source tab.

This includes information about the processing steps that were applied to the source image. You will find it listed under the Processing History section.

Symbolize and examine the calibrated imagery

You will now change the default symbology color scheme of the calibrated SAR image to better visualize the measured backscatter. Then you’ll examine the image and view the ships and other features it contains.

1. In the Contents pane, ensure that the Panama SAR - Radiometric Calibration.crf layer is selected.

   

2. On the ribbon, on the Raster Layer tab, in the Rendering group, click the Symbology button.

   

   The Symbology pane for the Panama SAR - Radiometric Calibration.crf layer appears.

3. In the Symbology pane, observe that the image is currently displayed as a Stretch, with a black-to-white color scheme.

   

   This means that all the pixels in the image are shown in black-to-white tones, with the lower measured backscatter values appearing in black, the higher values in white, with a gray gradient in between for the middle values.

4. In the Symbology pane, set the following parameters:
   • For Primary symbology, confirm Stretch is selected.
   • For Band, confirm the VV polarization is selected.
   • For Number of standard deviations, type 0.5.
   • For Gamma, type 2.0.

   

5. For Color scheme, expand the drop-down list and check the Show names box. Select the Viridis color scheme.

   

   On the map, the layer updates to the new symbology.

6. On the ribbon, on the Map tab, in the Navigate group, click Bookmarks and choose Northern Entrance.

   

   The map updates to the position specified by the bookmark, showing the northern entrance of the Panama Canal where ships are waiting to cross the Panama Canal.

   

   The areas of highest backscatter values appear in yellow and lowest backscatter values in indigo tones. The ocean appears mainly in indigo tones. The land mainly appears in blue tones. In the ocean, the ships appear yellow.

   In that image, the highest backscatter values correspond to a type of scattering named double-bounce. In this type of scattering, the radar signal reflects once off a vertical target (the ship) onto a smooth surface (the ocean) and reflects a second time off the smooth surface back toward the sensor. This scattering type often takes place on human-made structures, such as ships, and results in very high backscatter values. See the following diagram for an illustration of double-bounce scattering:

   

   Illustration of double-bounce scattering.

   Note:

   Learn more about double-bounce scattering in the Types of Scattering section of the Fundamentals of Synthetic Aperture Radar (SAR) guide.

7. On the ribbon, click Bookmarks again and choose the Southern Entrance bookmark to view the ships waiting to cross the Panama Canal at the south entrance.

   

8. On the map, zoom in and out with the mouse wheel button and pan to explore the image.

   Most ships appear as small elongated yellow shapes. Since ships give off double-bounce scattering, the backscatter contains high values that can sometimes be interpreted as bright plus signs or stars. You will examine two examples.

9. On the ribbon, click Bookmarks again and choose the Ships bookmark.

   

   It's important to note that the crossing lines of the plus signs and stars do not represent the ship, but are simply an artifact of SAR imaging.

10. Zoom out and spend more time observing the different features in the image, such as the coastlines, land, and water bodies.
11. When you are done examining the image, on the ribbon, click Bookmarks again and choose the Overview bookmark to go back to the full extent.

    

    You’ll save your project.

12. On the Quick Access Toolbar, click Save to save your project.

    

    So far in this workflow, you set up the project, applied radiometric calibration to a SAR GRD image representing the Panama Canal and its surroundings, symbolized the image, and examined the ships and other features it contains. Next, you will start the analysis.

Analyze SAR data

The SAR image is now ready for analysis. First, you will detect the ships in the image. Then you’ll compare the ships detected to the original image and understand the importance of geometric terrain correction.

Detect Bright Ocean Objects

You will detect the ships in the SAR image using the Detect Bright Ocean Objects tool. The tool takes a gamma nought calibrated SAR image as input, and looks for clusters of high backscatter value pixels (that is, bright objects). It then filters them to keep only the bright objects of specified minimum and maximum width and length. The output is a feature class that shows rectangular boxes around the detected objects. The bright objects that will be detected in this image are ships. In other extents, they could also be different types of human-made features, such as oil rig platforms.

The tool uses two additional input layers: a feature class representing land polygons and a digital elevation model (DEM) raster. You will now add these layers to your map and understand their role.

1. At the bottom of the Symbology pane, click the Catalog tab.

   

2. In the Catalog pane, expand the Databases folder and the Detect_ships_with_SAR_imagery.gdb geodatabase, right-click the Land_Polygons feature class, and choose Add To Current Map.

   

   The Land_Polygons layer appears on the map.

   

   Note:

   The color is assigned at random and may vary.

   This is a polygon layer representing the landforms in the region. The Detect Bright Ocean Objects tool will use it as a mask so that it only tries to detect ships in the ocean and not on land.

   You’ll now add the DEM layer.

3. In the Catalog pane, browse to Folders > Detect_ships_with_SAR_imagery > Input Data, and expand the DEM folder. Right-click Panama DEM 30m.tif, and choose Add To Current Map.

   

   The DEM layer appears on the map. The Land_Polygons layer displays above it and partially obstructs it, so you’ll turn it off.

4. In the Contents pane, uncheck the box next to the Land_Polygons layer to turn it off.

   

   You can now see the DEM layer fully.

   

   The DEM layer is a raster that provides elevation data about the region. By default, it displays with a black-to-white color scheme in which the lowest elevation values appear in black and the highest ones in white. It will be used by the Detect Bright Ocean Objects tool to orthorectify the output. Orthorectification is the process of correcting apparent changes in the position of ground objects caused by the perspective of the sensor view angle and topographic relief.

5. In the Contents pane, turn off the Panama DEM 30m.tif layer, as you no longer need to visualize it.

   

   You will now run the Detect Bright Ocean Objects tool.

6. In the Geoprocessing pane, click the Back button.

   

7. On the Toolbox tab, under the Image Analyst Tools toolbox, in the Synthetic Aperture Radar toolset, click the Detect Bright Ocean Objects tool to open it.

   

8. Set the following tool parameters:
   • For Input Radar Data, choose Panama SAR - Radiometric Calibration.crf.
   • For Output Feature Class, type Detected_Ships.
   • For Output Type, confirm that Bounding box is selected.

   

   Note:

   The Detect Bright Ocean Objects tool requires SAR input that was calibrated with the Gamma nought method (see the Apply radiometric calibration section earlier in the tutorial). Another requirement is that the SAR input should not have undergone any other processes, such as despeckle or geometric terrain correction, because this would change the pixel distribution needed to detect bright objects.

9. Set the following width and length parameters:
   • For Minimum Object Width, type 5.
   • For Maximum Object Width, type 500.
   • For Minimum Object Length, type 10.
   • For Maximum Object Length, type 700.

   

   The tool will only keep the detected objects that are between 5 and 500 meters in width and between 10 and 700 meters in length.

10. Specify the land polygon and DEM layers that the tool will use.
    • For Mask Features, choose Land_Polygons.
    • For Feature Type, choose Land polygon.
    • For DEM Raster, choose Panama DEM 30m.tif.
    • Confirm that Apply Geoid Correction is checked.
    • For Mask Tolerance, type 300.

    

    Note:

    To learn more about any tool parameter, point to it and click the information button.

    

11. Click Run.

    After a couple of minutes, the Detected_Ships layer appears on the map.

Symbolize and examine the detected ships layer

You’ll change the symbology of the Detected_Ships layer and examine the results.

1. In the Contents pane, ensure the Detected_Ships layer is selected.

   

2. On the ribbon, on the Feature Layer tab, in the Drawing group, click the Symbology button.

   

3. In the Symbology pane, click the symbol tile.

   

4. On the Properties tab, under Appearance, for Color, expand the color palette and choose No color.

   

5. For Outline color, expand the color palette and choose a bright orange, such as Electron Gold.

   

6. For Outline width, choose 2 pt.

   

7. Click Apply.

   On the map, the detected ships appear as bright orange rectangles displaying over the Panama SAR - Radiometric Calibration.crf layer.

8. Zoom in and pan to examine the results.

   

   Most ships were successfully detected. However, you might notice that there is a slight discrepancy between the orange boxes and the location of the ships on the SAR image.

   

   You will understand and address this issue in the next section.

Apply Geometric Terrain Correction

The Detect Bright Ocean Objects tool produced an orthorectified output (with the help of the DEM), whereas the Panama SAR - Radiometric Calibration.crf image has not yet been orthorectified. This explains why there is a discrepancy between the orange boxes and the location of the ships on the SAR image. To enable a more accurate comparison, you will now orthorectify the SAR image using the Apply Geometric Terrain Correction tool.

1. In the Geoprocessing pane, click the Back button.
2. On the Toolboxes tab, under the Image Analyst Tools toolbox, in the Synthetic Aperture Radar toolset, click the Apply Geometric Terrain Correction tool to open it.

   

3. Set the following tool parameters:
   • For Input Radar Data, choose Panama SAR - Radiometric Calibration.crf.
   • For Output Radar Data, type Panama SAR - Geometric Terrain Correction.crf.
   • For Polarization Bands, confirm that the VV box is checked.
   • For DEM Raster, select Panama DEM 30m.tif.
   • Confirm that the Apply geoid correction box is checked.

   

4. Click Run.

   After several minutes, the new layer appears. You’ll change its symbology.

5. In the Contents pane, ensure that the Panama SAR - Geometric Terrain Correction.crf layer is selected.
6. On the ribbon, on the Raster Layer tab, in the Rendering group, click Symbology.
7. In the Symbology pane, set the following parameters:
   • For Color scheme, expand the drop-down list, check the Show names box, and select the Viridis color scheme.
   • For Number of standard deviations, type 0.5.
   • For Gamma, type 2.0.

   

   You’ll visualize the ships located at the northern entrance.

8. On the ribbon, on the Map tab, click Bookmarks, and choose the Northern Entrance bookmark.
9. Zoom in and pan to observe the detected ships.

   

   The Detected_Ships feature class now aligns with the ships in the Panama SAR - Geometric Terrain Correction.crf raster layer.

10. In the Contents pane, turn the Panama SAR - Geometric Terrain Correction.crf layer off and on to see the difference of position.

    

    The difference in alignment is clearly visible.

    

    SAR layers without geometric terrain correction (left) and with geometric terrain correction (right) with a focus on detected ships.

    Note:

    Some of the larger ships correspond to multiple polygons, because different parts of the ship were detected separately. To obtain a more precise ship count, you can optionally use the Pairwise Dissolve tool, making sure to uncheck the Create multipart features box.

    To better understand the notion of orthorectification, you will look at other areas of the image.

11. On the map, zoom out and pan to a location where you can see the coastline.

    

12. In the Contents pane, turn on the Land_Polygons layer.

    

13. Turn the Panama SAR - Geometric Terrain Correction.crf layer off and on to see how the coastline’s position has also been corrected.

    

    SAR layers without geometric terrain correction (left) and with geometric terrain correction (right) with a focus on coastlines.

    You can observe that the orthorectification process happened throughout the entire image.

14. On the ribbon, on the Map tab, click Bookmarks, and choose the Overview bookmark to go back to the full extent.
15. In the Contents pane, turn off Land_Polygons and ensure that Panama SAR - Geometric Terrain Correction.crf is turned on.

    

16. Press Ctrl+S to save your project.

In this part of the workflow, you detected the ships in the SAR image with the Detect Bright Ocean Objects tool. Then, after applying geometric terrain correction to the SAR image, you compared the ships detected to the image.

Generate a vessel density map

You will now create a vessel density layer to indicate where the detected ships are concentrated. Then, you’ll symbolize the layer and interpret the results.

Summarize Within

You will create the vessel density layer with the Summarize Within tool and a 5-by-5-kilometer grid. First, you’ll add the grid layer to the map.

1. In the Catalog pane, expand Databases and Detect_ships_with_SAR_imagery.gdb, right-click Grid_5km, and choose Add To Current Map.

   

   The grid appears on the map. Each of its cells measures 5 by 5 kilometers.

   

   For now, the grid doesn’t contain data.

   Note:

   This empty grid was created with the Create Fishnet tool.

   You will use the Summarize Within tool to count how many ships are in each cell of the grid.

2. In the Geoprocessing pane, click the Back button. Type Summarize Within in the search box and click Summarize Within (Analysis Tools) to open it.

   

   Note:

   Summarize Within (GeoAnalytics Desktop Tools) is a different tool and is not used in this workflow.

3. Set the following tool parameters:
   • For Input Polygons, select Grid_5km.
   • For Input Summary Features, select Detected_Ships.
   • For Output Feature Class, accept the default name.
   • Uncheck the Keep all input polygons box.
   • For Shape Unit, select Square meters.

   

4. Click Run.

   The new layer appears on the map, but you can’t distinguish it from the original grid layer.

5. In the Contents pane, turn off the Grid_5km layer.

   

   You can now observe that, in the new layer, only the cells that had one or more ships were retained.

   

   You will look at the attribute table for this layer.

6. In the Contents pane, right-click Grid_5km_SummarizeWithin and choose Attribute Table.

   

7. Examine the content of the Grid_5km_SummarizeWithin attribute table.

   The Count of Polygons attribute was added by the Summarize Within tool and contains the number of ships for each grid cell.

   

8. When you are done examining the attribute table, close the Grid_5km_SummarizeWithin attribute table window.

   

Symbolize and interpret the vessel density layer

You will symbolize the vessel density layer and interpret the results. First, you'll give the layer a more meaningful name.

1. In the Contents pane, click the Grid_5km_SummarizeWithin layer to select it and click it again to rename it. Type Number of ships per grid cell and press Enter.

   

   Next, you’ll change the symbology to reflect the number of ships per grid cell.

2. Ensure that the Number of ships per grid cell layer is selected.
3. On the ribbon, on the Feature Layer tab, in the Drawing group, click Symbology.
4. In the Symbology pane, for Primary symbology, choose Graduated Colors.

   

5. For Field, select Count of Polygons.

   

   Now that you’ve symbolized the vessel density map, you’ll look to see where there is the most ship congestion.

6. On the map, examine the Number of ships per grid cell layer, as well as the corresponding legend in the Contents pane.

   

   The areas with the highest density appear in red, and those with the least density appear in yellow. You see that most ships are aligned with the northern and southern entrances of the canal and cover about a 75-square-kilometer area at each entrance.

7. Press Ctrl+S to save your project.

In this tutorial, as an analyst for the Panama Traffic Services agency, you used SAR imagery to determine maritime traffic congestion in the Panama Canal region. You applied radiometric calibration to a SAR image and changed its symbology. You then detected the ships in the SAR image with the Detect Bright Ocean Objects tool. After applying geometric terrain correction to the SAR image, you compared the ships detected to the image. Finally, you generated a vessel density layer to indicate where the detected ships are concentrated, and you symbolized the layer and interpreted the results.

An organization such as the Panama Traffic Services agency aims to minimize ship congestion and ensure smooth operations by monitoring various factors and implementing efficient management strategies. Note that in a real-life setting, a vessel density map would be only one of the tools used to reach these goals. Examples of other important factors are weather conditions, the volume of incoming and outgoing cargo, and the ports’ cargo handling capacity.

You can find more tutorials in the tutorial gallery.

Acknowledgements

• Panama SAR GRD image (ICEYE_GRD_SC_244270_20220207T192226.tif): ICEYE data [2021-2022]. Retrieved from ICEYE's website, processed by Esri.
• Land_Polygons: Land polygons data derived from OpenStreetMap and retrieved from OpenStreetMap Deutschland.
• Panama DEM 30m.tif: Copernicus GLO-90 Digital Elevation Model from the European Space Agency, distributed by Open Topography.
• Grid_5km: Layer was created with the Create Fishnet tool.
• World Topographic Map sources: Esri, HERE, Garmin, FAO, NOAA, USGS, OpenStreetMap contributors, and the GIS User Community
• World Hillshade layer sources: Esri, Airbus DS, USGS, NGA, NASA, CGIAR, N Robinson, NCEAS, NLS, OS, NMA, Geodatastyrelsen, Rijkswaterstaat, GSA, Geoland, FEMA, Intermap, and the GIS user community