People have always been fascinated by views of our planet from vantage points other than our own eyes. Artists and photographers have trekked to remote lands or ascended to great heights by balloon, by airplane, and by spacecraft, to show our planet as we cannot see it from the surface. We now have a new source of art and wonder provided by remote sensing satellites, showing us that earth can be seen as art.

A number of books have been published that contain images of earth from space, and of course there is a large number of web sites that contain such images. Yet, these mostly have left me unsatisfied. Almost all of these images are oriented towards showing us specific places ("Gee, there's the Grand Canyon. Look how small it is from space!") with unimaginative, although realistic, colors schemes. Ansel Adams demonstrated, by capturing light and shadows generated by the natural wonders of our world, that these places are also magnificent artistic wonders. Part of what makes his photographs great art is that in addition to showing us places, they also show us abstract compositions and shapes that could have been created by a great painter. My goal is to use remote sensing satellite data to show our planet in ways that our eyes are incapable of seeing, and as a result, to reveal the abstract compositions, shapes, and colors of earth that are hidden in this data.

Remote sensing satellites do not take photographs like ordinary film or digital cameras. Instead, they capture the reflected energy at each of several narrow bands of wavelengths of light, including visible and infra-red. For example, the Landsat satellites record six different bands including blue-green, red, near infra-red, and short-wavelength infra-red. NASA's airborne AVIRIS instrument records 224 bands across this spectrum. Different surface features reflect these wavelengths at different intensities with most of this differentiation occurring among the infra-red bands. However, the raw data produced by these satellites require significant image processing to produce interesting images.

I first became interested in remote sensing data as an application of some statistical theory I had been developing. As the speed and capability of PC's improved, it became feasible to perform significant image processing on large data sets using relatively inexpensive PC's rather than supercomputers. This enabled me to experiment with large remote sensing data sets using special image processing routines I had developed. It quickly became apparent that the images produced by my routines had considerable artistic merit beyond their scientific use. The goal of my image processing is to create color assignments from this data that emphasize the composition, shapes, and drama of the scenes I have selected.

Please note that these images are not taken from any other web site, and they are not photographs taken from space that have been reprocessed. Photographs cannot capture the detail, nor can they differentiate among different features, like the sensors on remote sensing satellites. For example, compare this photograph taken by astronauts aboard the Space Shuttle of the Anti-Atlas mountains of Morocco with my image of this region. Here is some of the detail present in that image.

The images on display here represent low-resolution versions of the image files used for making prints. The description page of each image has a link to a crop of the image file that shows the detail present in the full image. These images represent a sampling of the images I have constructed thus far, and I am continuing to search for additional interesting data sets. If you have a special request for an image of a particular area, please contact me concerning availability of data for that area.

The Earth is a beautiful oasis in a vast, mostly empty, universe. My goal with this work is to increase appreciation for the unique beauty of our small planet. I hope I have succeeded.

Larry P. Ammann
Professor of Statistics
University of Texas at Dallas