TerraLook: Satellite Imagery to View a Changing World
Landsat and other satellite sensors have been collecting images of Earth from space for more than 35 years. For the most part, use of this data has been limited to the scientific and land management communities. TerraLook was developed to broaden the population of satellite image users by providing both ASTER and Landsat data as simulated natural-color JPEG images. TerraLook aims to serve user communities who have a need for images of the Earth but do not have technical remote sensing expertise or access to expensive and specialized scientific image processing software. TerraLook Collections consist of readily usable remotely sensed images designed for visual interpretation and are distributed at no cost to the user. TerraLook collections will be of value to anyone who wants to see the changes to Earth’s surface since 1972. Users who have a need to perform digital analysis can acquire the source data. The georeferenced TerraLook product is compatible with most GIS and web mapping applications.
The TerraLook product and software build on the success of the NASA Jet Propulsion Laboratory’s Protected Area Archive project. Originally developed for the conservation and emergency response communities, TerraLook has received support from The Nature Conservancy, the World Bank Development Grant Facility Program’s Inter-American Biodiversity Information Network Connectivity Project, the U.S. State Department, the Gordon and Betty Moore Foundation’s Amazon Protected Area Network project, and The World Conservation Union.
In 2007, USGS EROS became the operational home for the TerraLook product. JPL continues to serve a major role in outreach, particularly to the conservation community, in the coordination of the development of the TerraLook software, and in the evolution of the TerraLook product. The TerraLook project is a joint project between USGS EROS and NASA JPL.
TerraLook collections consist of images derived from two satellite imaging programs:
Since 1972, a series of Landsat satellites has been collecting digital images of Earth. Visit the NASA Landsat and USGS Landsat Web sites for further information about the Landsat Program. The Landsat Global Land Survey (GLS) images were selected from 2.5 million images in the Landsat archive to provide global coverage of the Earth’s surface at four epochs of time (circa 1975, 1990, 2000, and 2005). Images for each epoch were selected for best cloud-free coverage during the local growing season. The nominal pixel size is 80 m for the 1975 images and 30 m for the 1990, 2000, and 2005 images. These images, captured by different types of Landsat sensors (MSS, TM, and ETM+), can be combined within collections to create a time series of images:
ASTER, one of five sensors on the Terra spacecraft, has captured over 1.5 million imaes since its launch in 1999. To learn more about ASTER, visit the ASTER Web site hosted at NASA JPL. TerraLook images can be derived from any image in this archive from early 2000 to the present to provide the most recent images captured at twice the resolution (15 meters) of the best Landsat sensors.
The TerraLook images are bundled with image footprint and metadata files that describe the images and the collection. The image footprint files delineate the image boundaries so the image location can be easily seen, and the metadata files contain variables that describe the sensor, acquisition time, and the name of each image file. The image footprints are stored as industry-standard GIS shapefiles. The image and collection metadata are stored in standards-compliant XML files. TerraLook products consisting of images, footprints, and metadata are bundled zip files.
A TerraLook collection can consist of any combination of images from these ASTER and Landsat GLS datasets. A common use is the study of Earth over time, however many users simply want an image of their area of interest. A typical choice is a collection through time for one or more cities or parks, or for a region or country, but a collection can also consist of images from one sensor for a larger area within the same time period. The selection of images is completely up to the user.
In addition to the custom collections you can create, Archived Collections from locations around the world have already been compiled, providing an easy way for you to get acquainted with the imagery without first having to make selections: merely download the zipped file.
TerraLook images have many advantages. The JPEGs are compressed about 12:1, making them quick to download. Simulated natural color makes the imagery easy to understand by a wide range of users. The algorithm used to process TerraLook images permits the creation of a visually consistent time series from the earliest Landsat sensor launched in 1972 to the currently orbiting ASTER and Landsat sensors.
TerraLook images are a simulated natural color rather than a "true" or "natural" color image. True- or natural-color images are created from blue, green, and red light. Most satellite sensors do not collect data in the blue spectral band of the electromagnetic spectrum. Even when a blue band is available, the blue light, as seen from space, is scattered by atmospheric moisture (the reason for blue skies) creating a very noisy blue band, particularly in humid areas. Most satellite sensors collect near-infrared (NIR) data, which is sensitive to the health of vegetation. In response to the lack of a blue band and the availability of an information-rich NIR band, most satellite images are viewed using some combination of visible and infrared data. TerraLook images use the bands derived from the red, green, and NIR bands collected by Landsat and ASTER sensors with the algorithm described below to form a simulated natural-color image.
Red = Red
Green = 2/3 Green + 1/3 NIR
Blue = 2/3 Green – 1/3 NIR
The synthetic green band is enhanced through the addition of information from the vegetation-sensitive NIR band. The synthetic blue is created from the spectrally-most-similar green band with the vegetation information suppressed.
TerraLook images are georeferenced with industry standard world files, which makes them suitable for use in most Geographic Information Systems, image processing systems, and Web mapping applications. Just as important, the images can be viewed in any software that can read and display JPEG images; this includes most Web browsers, word processors, and graphics packages. Pixel size is in decimal degree equivalent at the center of the image of the source image or approximately 111,200 meters per degree at the equator. GLS 1975 has 60 meter (0.00054 degrees latitude) pixels; GLS 1990, 2000, and 2005 have 30 meter (0.000270 degrees latitude) pixels; and ASTER has 15 meter (0.000135 degrees latitude) pixels. Pixel size along longitude can be roughly estimated by pixel size along the latitude / cos(latitude) at the image center. As a consequence, TerraLook image pixels are not square. To display correctly in image display applications that cannot handle rectangular pixels, TerraLook images must be stretched by the ratio of the x and y pixel sizes.
TerraLook Software is a bilingual (Spanish and English), open-source product designed for use with TerraLook Collections. The easy-to-use software allows you to view, compare, and annotate TerraLook imagery. TerraLook Software displays the image footprints, allows images to be viewed by selecting an image footprint, and manages each Landsat epoch and ASTER archive separately, making it easy to study Earth through time and to construct virtual mosaics. The images can be annotated to create graphics for meeting or classroom presentations, and vector layers can be overlaid on the images (a vector layer can be overlaid on top of an image to show a point location, a line such as a road, or a polygon such as country boundary). Some vector layers come with the images, and additional vector layers can be created interactively to store interpreted information such as park or urban boundaries, to insert locations of ground-sampled information, or to compare changes in surface features over time. A great number of vector layers are also available via the World Wide Web.