Difference between revisions of "Creating A Digital Elevation Model using Sentinel-1 Data"
Alex Fortin (talk | contribs) |
Alex Fortin (talk | contribs) |
||
| Line 106: | Line 106: | ||
[[File:PreProcessing_manifestsafe |thumb|none|frame| Select and open the manifest.safe file]] |
[[File:PreProcessing_manifestsafe |thumb|none|frame| Select and open the manifest.safe file]] |
||
| − | * The file should open in the product explorer window in the top right |
+ | * The file should open in the product explorer window in the top right<br> |
| − | * Repeat these steps for the second image.<br> |
||
=== S-1 TOPS Split === |
=== S-1 TOPS Split === |
||
Revision as of 16:27, 7 April 2025
Contents
Guide to Creating a Digital Elevation Model Using Sentinel-1 Data in the Sentinel Application Platform (SNAP)
Introduction
Digital Elevation Models (DEMs) are and essential tool in geospatial analysis. They provide detailed representations of the earth's surface elevation and are widely used in application like, hydrological modeling, land cover classification, infrastructure planning, and environmental monitoring. Synthetic Aperture Radar (SAR) data, such as that provided by Sentinel-1, offers a method to create DEMs.
This tutorial will guide you through the steps to create a Digital Elevation Model using Sentinel-1 data in the European Space Agency's (ESA) Sentinel Application Platform (SNAP). SNAP, the official open-source software developed by the ESA, provides an extensive range of tools for processing Sentinel data, including SAR interferometry (InSAR). Sentinel-1, a part of ESA's Copernicus Program, provides high-resolution C-band SAR imagery that can be processed using interferometric techniques to extract elevation data.
This guide will allow users to:
- Acquire and prepare Sentinel-1 SLC (Single Look Complex) data.
- Perform coregistration of SAR images.
- Generate an interferogram and apply phase filtering.
- Unwrap the phase and convert it into elevation values
- Export and visualize the final DEM
No experience with SAR data is required for this tutorial, but a basic understanding of remote sensing will be helpful. This tutorial caters to students, researchers, or other GIS professional in need of Digital Elevation Models for areas they are studying, and will equip you with the base skills needed to create and analyze these models using radar imagery in Snap
Data Consideration
Interferometric SAR or InSAR is a form of geospatial analysis that uses the measurement of phase from two or more synthetic aperture radar (SAR) images. SAR imagery is a form of active remote sensing where the sensor emits microwaves towards the target and receives the reflected waves. Several factors can influence the accuracy of InSAR analysis, and the table below summarizes these factors, their effects, and some best practices to create the best DEM possible.
The way in which these waves are reflected are called scattering mechanisms. In general, these mechanisms can create coherent or incoherent scatter. Coherent scatter is caused by built structures, calm water, or urban areas. Incoherent scatter is caused by forests, agriculture, and rough water. In InSAR analysis, coherence has a big effect on accuracy, meaning that it works best in areas with coherent scatter.
| Factor | Characteristics | Effect on Analysis | Recommendation |
|---|---|---|---|
| Land Cover | Urban landscapes, forests, built structures, agriculture, water | Forests, agriculture, and rough water cause incoherent scatter which negatively impacts the accuracy of analysis.
Urban landscapes, built structures and calm water cause coherent scatter which is desirable for InSAR analysis |
For best results, choose a study area with land cover that causes coherent scatter |
| Image Selection | Orbit of satellite, viewing geometry, path of sensor | Any differences in the extent of the selected images, the viewing geometry, the path of the sensor (ascending or descending) will have significant repercussions on the coherence between the images and the accuracy of the DEM | Ensure that the selected images cover the same extent and have the same path. If not, the DEM will not be usable.
For best results, select images with as similar viewing geometry as possible to reduce incoherence |
| Time | Period between the capture of images | The longer the time between image captures, the greater the incoherence between the images and the lower the accuracy of your DEM | For best results, choose images with the lowest possible time between captures. For Sentinel-1 data in Canada, this is usually 12 days |
| Wavelength | L-band (~20 cm), C-band (~5 cm), X-band (~3 cm) | Shorter wavelengths get reflected by smaller features, meaning that C- and X-band get reflected by tree canopies and other vegetation, while L-band passes through | For best results, use the longest wavelength available to you. If you are using C- or X-band SAR, recognize that your output is a hybrid DEM and DSM depending on the land cover in your study area |
| Date | Date at which images were captured | If a significant event that impacts the landscape of the study area occurred around the time the images were captured (ie. earthquake, volcanic eruption, bombing, flood) the DEM will not be accurate. | For best results, check if any major geological or anthropogenic occurred around the dates of image capture. If they did, change your selection |
Prerequisites
Materials List
1. Sentinel-1 Toolbox (SNAP 11)
2. Statistical-cost, Network-flow, Algorithm for Phase Unwrapping (SNAPHU V2.0.7)
3. Copernicus Browser (Link to browser)
Copernicus
Downloading Sentinel-1 Data
Access the Copernicus Browser
- Select "Login" on the pop up, as you are required to sign in or create a free account to download data from their platform.
- On the sign in page either sign in with your credentials or click the “Register” button and follow the instructions to create your account.
- After logging in, navigate to your AOI (Area Of Interest) on the global map. Remember DEMs created from SAR data work best in urban or non-vegetated areas (Add a little bit of info)
- Once at your AOI, select the "SEARCH" tab on the upper left side of the screen (Red Outline Below).
- Once this is done, select both the Sentinel-1 and Level-1 SLC Boxes (Red Outline Below).
- Next, Scroll down and set the time range using the "From" and "Until" drop down calendars (Red Outline Below). An appropriate range should be within a few years of your analysis (e.g. this tutorial made in early 2025 uses images from 2022). If significant changes in the landscape of you AOI have happened (e.g. earthquake, volcano eruption, flooding), ensure that the time range is set after this event. (Minimum time range...)
- Click the "Search" button at the bottom of the menu. This will generate results that satisfy your criteria. The menu will display a list of images captured in the time range and the map will show outlines of the image extents in blue.
- You can now select an image by clicking on a blue extent that covers your AOI. A pop up window should appear listing the images with that coverage within your time range (See image below). For InSAR, we require two images. Remember that a low temporal baseline (Time between images) will increase the accuracy of you DEM, so selecting two images with the shortest period between captures is suggested.
- To download the images click the icon on the far right of each row (Red Outline seen above).
- Once the zip folders containing the images have downloaded, extract the files to the folder you will be using as your directory in the next steps.
SNAP 11
Preprocessing (To be completed with both images)
Opening Data in SNAP
- After launching SNAP, click the folder icon in the top left of the window (Red Outline) or navigate to File -> Open Product. This will open a pop up window.
- Use the "Look in" dropdown menu (Outlined in red below) to navigate to your directory folder. Once this is done double click the .SAFE folders (highlighted in yellow below) to open them in the file browser.
- Once in the folder select the "manifest.safe" file and click open.
- The file should open in the product explorer window in the top right
S-1 TOPS Split
The S-1 TOPS Split tool allows the user to select which part of the image to use in the analysis. Sentinel-1 images are captured in subswaths and burts. The subswaths divide the extent horizontally in three (3) zones. These subswaths are then divided into vertical bursts. While we are limited to selecting one (1) subswath, we can select multiple bursts, as long as they are adjacent.
- In the I/O parameters tab, make sure the correct image is selected as source, and verify that you are saving to the proper directory.
- In the Processing Parameters tab, select the subswath and burst(s) that cover you AOI. The subswath is selected from the dropdown subswath menu, and burst(s) are selected by dragging the arrows (outlined in red) along the bar. The outline of the area of coverage will be shown on the map below in a white outline. At this stage you also select the polarization of the image. Sentinel 1 SAR data comes in HH, VH, and VH polarization, each with its own uses.
Apply Orbit file
This tool applies information about the orbit of the sensor to the image, correcting for drift in the satellite's path and ensuring that the image is accurate.
- Click Radar -> Apply Orbit File
- In the I/O Parameters, make sure the file ending in "_split" is the source
Back Geocoding
This tool combines the input images into a single image.
- Click Radar -> Coregistration -> S-1 TOPS Coregistration -> Back-Geocoding
Enhanced Spectral Diversity (Only required if using multiple bursts)
This tool corrects for the Doppler effect which occurs across bursts due to the motion of the satellite during data collection. This process increases accuracy and creates continuity across bursts.
- Click Radar -> Coregistration -> S1 TOPS Coregistration -> S-1 Enhanced Spectral Diversity
- I/O Parameters: make sure the "_split_Orb_Stack" ending file is the Source Product and run.
Interferogram Formation
Interferograms allow for the visualization of phase difference between the two input images.
- Click Radar -> Interferometric -> Products -> Interferogram Formation
- In the I/O Parameters tab, make sure the file ending with "_Orb_stack_esd" is the Source Product
- In the Processing Parameters tab, check Subtract Flat-Earth Phase (marked in red) and Substract Topographic Phase (marked in blue) boxes. The Subtract Flat-Earth Phase function accounts for curvature of the earth, and the Subtract Topographic Phase function uses DEM to reduce impact of topographic variation on accuracy.
- Select SRTM 1Sec HGT (Auto Download) as te Digital Elevation Model (highlighted in yellow)
TOPs Deburst (Only required if using multiple bursts)
This tool removes the line between bursts, merging the bursts into a continuous image
- Click Radar -> Sentinel-1 TOPS -> S-1 TOPS Deburst
- In the I/O Parameters tab, ensure that the file ending with "_Stack_esd_ifg" is the Source Product
- Run the tool
Phase Filtering
Phase Filtering reduces phase noise, this improves the signal noise ratio of the image resulting in a clearer image.
- Click Radar -> Interferometric -> Filtering -> Goldstein Phase Filtering
- In the I/O Parameters tab, ensure the file ending with “esd_ifg_deb” is set as the Source Product and run the tool.
SNAPHU
The DEM Product
Details
- Last Updated: 2025-04-07
- Published on: 2025-04-07
- Created By: Alex Fortin, Alex Hoferek, Leo Keenan
