The following is a brief tutorial showing a step by step procedure for extracting a digital surface model (DSM) from stereo imagery and converting to a digital terrain model (DTM). In this tutorial, Pleiades 1A imagery was used, however this workflow is similar for DEM extraction from all stereo optical data.
A DSM (also referred to as a DEM) extracted from stereo images represents the earth’s surface and includes all objects on it, for examples, buildings and trees. Many applications require a DTM which represents the bare ground surface without any objects. To convert a DSM to a DTM through manual editing is a very time consuming process. An automatic DSM to DTM conversion program was developed at PCI Geomatics.
The data used in this tutorial is a sample Pleiades primary data set consisting of panchromatic, multispectral, PMS and tri-stereoscopy images of Melbourne, Australia: https://www.intelligence-airbusds.com/en/23-sample-imagery . A great innovation of the Pleiades system is to offer high resolution stereoscopic coverage capability. The stereoscopic coverage is realized by only a single flyby of the area, which enables collection of a homogeneous product quickly. In addition to the “classical” forward and backward looking stereoscopic imaging, Pleiades can acquire an additional quasi- vertical image (tri-stereoscopy), thus enabling the user to have an image and its stereoscopic environment. In general, a forward and backward looking stereo pair produces the highest accuracy, but this combination's use is limited to areas with gentle terrain. A nadir and forward/backward looking stereo pair can be used in most types of terrain.
Initial Project Setup
1. Open the Geomatica OrthoEngine application.
2. Click File > New
3. Give your project a Filename, Name and Description
- Select Optical Satellite Modeling as the Math Modeling Method
- Select Rational Function (Extract from image) under Options
- Click OK
4. Input the appropriate Output projection and GCP projection information for your project
5. Select Data Input as the Processing step
The data to input will be the back scan image and the forward scan image. When viewed in Focus the images will look similar to the images below.
- Click Open a new or existing image
- Click New Image
- Navigate to the location of the data. Select the DIM_PHR1A_XXXXXXXXXX.xml file in the image folder
- Select Yes when asked if you want to import the data file to a .pix file for optimized processing
- Select a file name and location for your output .pix file
- Select Yes when asked if you want to create overviews now
- Repeat step 5b to 5f to add all images to the project
6. To view the output image select the image from the Open Image window and click Quick Open. Close both windows and move on to step 7
You can optionally choose to collect GCPs and TPs for your project to improve the math model. More information on how to collect GCPs and TPs is outlined in the optical orthorectification tutorial: https://support.pcigeomatics.com/hc/en-us/articles/209885893-Satellite-Orthorectification-Workflow#h_52968271361535391309828
For DEM Extraction you want to ensure that the stereo images are very well aligned to each other to ensure the highest quality DEM. Therefore it is highly recommend to at minimum collect TPs to tie the images together.
7. On the OrthoEngine toolbar select DEM From Stereo as the Processing step
- Click Create Epipolar Image
- Select left and right image
- With both images selected, click the Add Epipolar Pairs To Table button
- Click Generate Pairs
- Click OK to the pop-up message that states the epipolar pairs completed successfully
- Close the Generate Epipolar Images panel
8. In the OrthoEngine toolbar select DEM From Stereo as the Processing step
- Click Extract DEM automatically
- Check the epipolar pair by checking the Select box associated with that record
Under DEM Extraction Options:
- Select SGM (Semi-global matching) as the Extraction method
There are two extraction methods available. NCC (Normalized cross-correlation) and SGM (Semi-global matching). SGM is based on newer technology and produces higher-quality results with fewer errors and higher detail, but processing time is increased. However, if lower-resolution DSMs are adequate, and faster processing time is preferable, consider selecting the NCC method. More information about these extraction methods is available in the Geomatica Help.
- Check the Ouput DEM vetical datum. By default this is set to Ellipsoid for Rational Functions projects. You can change the vertical datum to MSL if you prefer to automatically convert the output DEM to MSL instead.
- Select a Pixel sampling interval of 2
- Check mark Epipolar tracking
The quality of the DSM extraction is highly dependent on how accurately the epipolar lines between stereo images are aligned. Errors in the epipolar alignment can vary across a stereo pair resulting in a DSM where there are large patches of poor quality elevation values. This is especially true when DSM’s are generated at full resolution. Errors can occur even if the epipolar lines are shifted by a single line.
The Epipolar Tracking option enables tracking of changes in the epipolar line over the stereo pair and can automatically compensate for small gradual errors.
Epipolar tracking increases processing time (typically by 20% to 30%) and there is a small possibility of introducing errors in an otherwise good DSM.
- Select Create Geocoded DEM
- Select an output file name and location
- Set the X and Y Resolution to 1 meter
- Specify your preferred Output option
- Use last value: Uses the last value to replace the pixel values in the overlapped area in the existing geocoded DEM by the pixel values of the geocoded DEM that is added to the file.
- Average: Replaces the pixel values in the overlapped area by the average pixel values between the existing geocoded DEM and the one that is added to the file.
- Highest score: Uses the highest score to replace the pixel values in the overlapped area by the pixel value with the highest correlation score between the existing geocoded DEM and the one that is added to the file.
- Blending: Uses the mosaicking method to mosaic the DEMs together with blending between DEMs. This is the default value.
- Merge (new with Geomatica Banff): Available when SGM (Semi-global matching) is selected for DEM Extraction Options, this option merges a set of geocoded DEMs from a stereo airphoto or satellite project into a single DEM. The merged DEM is typically of higher quality, because each geocoded DEM has a different viewing geometry. Occlusion in one DEM can be filled from another DEM. In addition, the multiple DEM elevations of the same ground pixel can be used to detect blunders. Averaging multiple good elevations helps to increase vertical accuracy.
Merge also attempts to straighten building edges, if applicable. That is, with Merge selected, you can select the Clean up building edges with filter size check box, and then in the box to the right, type or select the size of the filter to use. The recommended value is 13.
- Click Extract DEM
Note: DEM creation will take a long time as these images have a high resolution. Occasionally you may get artifacts in water bodies. This is normal as the algorithm processes the DEM using smaller tiles. These artifacts would need to be edited manually using our DEM Editing Tools.
Pleiades Melbourne extracted DSM
Once the DEM is extracted you can choose to convert the DEM to a Digital Terrain Model (DTM). DTMs only includes the elevation of the ‘bare earth’ with vegetation, buildings and other man-made features removed (though roads and bridges are typically retained). A DTM is the preferred product to use in an orthorectification workflow. For more information on DSM to DTM conversion please see this tutorial: support.pcigeomatics.com/hc/en-us/articles/360015130032-DSM-to-DTM-Conversion