Space Imaging is becoming well known for its IKONOS images in the news, including pictures from space of the E3 surveillance plane, and more recently the World Trade Center and Pentagon. Few realize, however that Space Imaging has made available for free download a sizable database of satellite imagery for the United States and for International locations. Although the database is somewhat sparse so that there may not be coverage for your exact area of interest, certain features of this database make it ideal for DEM overlays. All the images are ortho rectified, aligned and georeferenced. Much of the free satellite imagery available in the past has been the opposite, making the overlay process difficult and inaccurate. Also, valuable information in the form of the sun azimuth and angle is supplied. Of equal importance, the scale and resolution of the images, unlike USGS DOQs are well-matched to the USGS 1:24000 DEMs.

Although the database is really meant as a marketing tool for Space Imaging, and the images are not large enough for professional use themselves, they are very suitable for the amateur cartographer. Moreover, they can give some insight into the use of this type of product for cartographic applications even for professionals. I will describe one way to apply the Space Imaging database to produce a satellite image overlay of a USGS 1:24,000 DEM. Here is the step-by-step procedure:

I looked for an area to map in my usual demonstration target, the vicinity of Glenn’s Falls, NY. (I use this as my standard mainly because it is ordinary.) I first selected an image by visiting the Eastview CartographicDatabase website. Images are chosen from their interactive menu by country. When you select a country you are presented with all the available images in that area. I chose a location around the town of Bolton Landing near Lake George, NY. The satellite image is shown at the upper right.

When you pull up the image, it will be accompanied by some useful metadata, including the latitude and longitude of the four corners of the image. Equally important, the image is ortho rectified, meaning it has been mathematically processed to remove distortion due to camera obliquity and is essentially a flat, corrected image aligned with north at the top. I copied down the coordinate data and downloaded the image. I imported it and then saved it as a .BMP using Paintshop Pro.

Next I visited www.gisdatadepot.com and downloaded the nearest 1:24,000 USGS DLG, which happened to be for Bolton Landing. I did this mainly to get the corner coordinates, because I know that these will match the coordinate data in the corresponding 1:24,000 DEM. (You can also extract them from the SDTS profile.) I then downloaded the Bolton Landing DEM.

3DEM is my best overlay tool, so I loaded the DEM into this application. Then I loaded the overlay image by choosing ‘operation’ and ‘apply map overlay’. I selected the ‘geo reference’ button. This brought up a text window that allowed me to drop the red and blue corner markers on the image and key in the corresponding latitude and longitude values. Since Space Imaging has conveniently supplied the coordinate information, this was a snap. I hit the ‘OK’ button and presto: the satellite image was placed in the correct position on top of my DEM. (Note that the match between the DEM and the image is not exact. I could have cropped the underlap out of the image by setting my viewing window but I wanted to illustrate the automatic registration of the image and so I left it unprocessed.)

I then rendered it as usual using the ‘operation’ and ‘3D view’ buttons. Despite the coordinate matching, the overlay alignment was not exact. I needed to adjust the position of the overlay slightly with respect to known ground features in order to improve the accuracy. Fortunately there was a prominent one, the Hudson River to the northwest of the map. 3DEM provides a nice feature for doing this called the ‘adjust map overlay’ selection under the ‘operation’ menu. Another trick is to take advantage of the sun azimuth and angle information supplied with the image. By matching these numbers in the ‘operation’ ‘modify 3D scene’ menus you can cast the lighting of the DEM exactly as existed when the image was taken. This means you will not have any conflicting terrain shadows to confuse the image.

The resulting maps are shown to the right. Color overlays would have made for prettier images, but the technical features of overlaying georeferenced satellite imagery over USGS DEMS should be apparent. I expect this will become more important with the growth of the IKONOS satellite image inventory.

As an exercise for the reader, see if you can identify the feature inside the white box on the image. It has a characteristic signature in many of the spectral bands used in satellite imagery and its unique shape makes it a useful ground identification feature.