Geomatica DEM Extraction

This section primarily deals with how to and issues related to extracting Digital Elevation Models in Ortho Engine. Some of the questions may also deal alternative solutions to perform tasks as in Focus or Modeler.

Geomatica Ortho Engine is a rigorous environment for generating a variety of imagery data products. You can create a digital elevation model (DEM) from stereo pairs of images, which are two or more images of the same area taken from different view points. This method can be very useful for creating a DEM for inaccessible areas. You can obtain stereo pairs from aerial photographs, digital or video images, and these sensors: ASAR, ASTER, IRS, IKONOS, SPOT, QUICKBIRD, and RADARSAT. Note: The DEM extraction process requires a special license and may not be available to you with your current Geomatica setup.

The process of generating a digital elevation model (DEM) consists of several steps:

• Convert the raw images into epipolar pairs.
• Extract epipolar DEMs from the overlap between the epipolar pairs.
• Geocode the epipolar DEMs and stitch them together to form one DEM.

What is an Epipolar Image?

How do I generate a DEM from ASTER data?

The Extract DEM Automatically Tool asks that I specify a minimum and maximum elevation. How does this affect the process?

When generating a DEM from multispectral imagery which band should I use? Or Does it matter?

What does the DEM Detail parameter mean?

What is the score channel?

I have created a DEM from my imagery but the output has lots of black squares across it, what are these and can it be fixed?

I do not like the results that were generated by the DEM process, does Geomatica have any tools that I can use to fix the output?

I generated several adjacent DEMs for my area, how can I create one DEM file?

Q: What is an Epipolar Image?

A) Epipolar images are stereo pairs that are reprojected so that the left and right images have a common orientation, and matching features between the images appear along a common x axis. Using epipolar images increases the speed of the correlation process and reduces the possibility of incorrect matches.
OrthoEngine uses image correlation to extract matching pixels from the two images and then uses the sensor geometry from the computed math model to calculate the x, y, and z positions of the interpolated DEM surface.

Back Top

Q: How do I generate a DEM from ASTER data?

A: A) The DEM extraction process ASTER data is similar to that of most of types of data and there is a tutorial that covers the process so see --- quick guide link –

Back Top

Q: The Extract DEM Automatically Tool asks that I specify a minimum and maximum elevation. How does this affect the process?

A) The approximate elevation helps the matching process by giving it a good first guess at the parallax that it should find between the images. The matching is based on a small tile (area) of imagery. If you don't give an estimated elevation, it has to assume an elevation of 0. If your ground elevation is significantly different, then the corresponding features from the right and left images might not appear in both of the small tiles, since the location has not been adjusted for the approximate offset due to elevation. So, setting the approximate elevation improves the success rate of the matching, and reduces errors.

Back Top

Q: When generating a DEM from multispectral imagery which band should I use? Or Does it matter?

A) It generally depends on the data. While it doesn't matter all that much which actual band you use, many people tend to avoid blue band, since that color band tends to saturate more. You should pick the band that offers the most detail, to improve the ability to correlate on features. By default OrthoEngine will use the first band; however you can change this default channel that is used when you are setting up your Epipolar pairs in the Generate Epipolar Images window. Scroll over to the Left & right Channels column and simply enter in the band that you wish to use in the extraction process.

Back Top

Q: What does the DEM Detail parameter mean?

A) The DEM extraction is hierarchical which means that it makes several correlation passes, starting at coarse resolution overview images, and working its way up the image pyramid. Setting the DEM Detail allows you to stop the correlation process at some point in the pyramid, so that you are using less detailed imagery for the correlation. For coarse data such as SPOT or RADARSAT, you would usually leave the detail at High. However, for high resolution imagery such as photos or IKONOS, the imagery will include features that you don't want in the DEM, such as individual trees, cars, people, sign posts, towers, etc. Often setting this parameter to Medium or Low will stop the correlation at a reduced resolution image, which often no longer includes small artifacts like this due to resampling. So, think about the types of features that you want to show up in your DEM. If you want the tiny high-res features, set Detail to high. If you want a more general, smoother DEM, set the detail to Low.

Back Top

Q: What is the score channel?

A) The Score channel is an auxiliary channel that is created for storing a scaled correlation score for each pixel in the DEM. This gives you a good start when you look at DEM editing in order to identify areas of possible errors where the overall correlation scores were not successful. Don't forget: if there are holes in the DEM where the correlation failed, they are often filled or interpolated in the DEM extraction process. Just because the pixels have values does not mean that they were successfully correlated on the ground. Use the score channel to get the true impression of where the data was extracted successfully.

Back Top

Q: I have created a DEM from my imagery but the output has lots of black squares across it, what are these and can it be fixed?

A) The output DEM will often contain black squares in your DEM, these features will usually contain your NODATA value and indicate that there was trouble interpolating the elevation for that pixel. If the holes are relatively small then you can use the DEM editing tools to generate a mask of the NODATA pixels and interpolate a DN value based on surrounding pixels.

Back Top

Q: I do not like the results that were generated by the DEM process, does Geomatica have any tools that I can use to fix the output?

A) Yes, Geomatica has DEM editing tools that can be used to edit existing or the DEMs that result from the interpolation process. You can also edit either the resultant geocoded DEM or edit the Epipolar DEMs that are used to produce the final raster surface prior to the interpolation surface.
If you want to edit the DEM before it is geocoded, do not select Create Geocoded DEM option. The DEM extraction will produce a file that contains the epipolar pair in the first channel, the correlation score (if selected) in the second channel, and the corresponding epipolar DEM in the third channel.
You can also edit the geocoded DEM after it is generated, however, the file will not include the epipolar pairs. If you selected Create Score Channel, the correlation score is saved as the first channel in the file and the geocoded DEM as the second one. You can then open the DEM editing tools window and edit the DEM from there.

Back Top

Q: I generated several adjacent DEMs for my area, how can I create one DEM file?

A) There are a few options available for this such as you could use Automatic Mosaicking to combine the DEMs, However using DEM from raster file tool is designed specifically for this purpose. (found under the Import and Build DEM processing step). This tool will allow you to import several raster DEM surfaces and then create one seamless (assuming that there is overlap between them) DEM output from them.

Back Top