NASA first produced the striking GEOCOVER positionally accurate orthorectified Landsat imagery covering the majority of the Earth's land mass in August of 1998. The Landsat scenes used to create the GEOCOVER-Ortho products were hand selected to provide the lowest cloud cover and highest quality data possible. GEOCOVER combined the best available horizontal and vertical control with detailed manually measured tie points between adjacent imagery to create a highly accurate spatial relationship between the raw Landsat imagery and the surface of the Earth resulting in a positional accuracy of 50 meters root mean square (RMS).
The GEOCOVER 2000 worldwide set of satellite imagery from the Landsat 7 Enhanced Thematic Mapper (ETM+) represents NASA's latest effort. The ETM+ 30m multi-spectral bands 2, 4, and 7 were pan sharpened with the 15m panchromatic band. These Landsat images were then ortho-rectified, georferenced, and projected in UTM WGS84.
The GEOCOVER data set has several attractive features
The images were selected to have exceptionally low cloud cover in most cases
The orthorectification and georeferencing are of exceptionally high quality, making them very suitable for mapping work
The images are already pan sharpened to 15m resolution
Landsat band 2 (green) band 4 (near infrared) and band 7 (medium infrared) are less susceptible to atmospheric scattering than the visible light bands 1, 2 and 3 (blue, green, red) resulting in virtually haze-free images
The images are stretched so that their dynamic range fully spans the 256 available digital number (DN) levels
One possible drawback for some applications is the image color. GEOCOVER images display with unnatural vegetation, and earth tones because of the selected bands as shown in the image below. Modifying these tones to make them appear more natural preserves all of the benefits of GEOCOVER while mitigating this drawback. PANCROMA has utilities for making such spectral transformations.
Orthorectified, pan-sharpened ETM+7/4/2 individual scenes are available from two sources: the NASA ZULU site and also the USGS Global Visualization Site (GLOVIS) at . The ZULU site is convenient but only offers the data in un-georeferenced MRSID format. This is not a bad thing itself. However MRSID is a highly compressed format capable of handling extremely large data sets. The ZULU tiles are edge matched and would be ideal for creating natural color mosaics. Unfortunately, when inflated each tile is in excess of 4GB in size, exceeding the limits of GeoTiff and even JPEG formats. The GLOVIS site archives tiles that are a more manageable size, but they are not edge matched. However, this data can be downloaded directly in GeoTiff format. In order to do so, select 'Collection | Landsat Legacy Collections | ETM+ Pan Mosaics". Select the scene of interest (New York City in the example above. Add it to the list for downloading and then download it to your computer.
Note that GEOCOVER images are considerably larger than pan sharpened Landsat ETM+ scenes. Typical image sizes are around 21,500 by 19,600 pixels equal to around 1.2GB, making processing of these files using the Windows operating system a bit of a challenge. You will need to shut down other foreground and background applications to process them successfully. After downloading and unzipping your file, you should have a single RGB GeoTiff format image. You can open it and inspect it by selecting 'File' | 'Open' to select your file and then selecting 'Display One File' | 'Display One Color Composite Image'. After inspecting, select 'Close Graphics Window and Reset'. The first step in transforming a GEOCOVER image is to decompose the RGB false color image into its corresponding band files. To do this select 'File' | 'Open' to select your file. You will get a PANCROMA warning message that the file may be too big for processing. Ignore the message. Select 'Display One File' | 'Display Landsat Geocover'. When the color image displays, select 'File' | 'Save Grayscale Image' | 'GeoTiff'. When prompted, supply a root file name. PANCROMA will save the three band files with a '2', '4' and '7' suffix, indicating the Landsat band designation.
The next step is to transform the band2, band4 and band7 files into a synthetic natural color RGB image. To do this, open the three band files successively, in the above order (i.e. band 2 first) by selecting 'File' | 'Open' in succession. Now select 'Landsat GEOCOVER 742/321 Spectral Transformer'. PANCROMA will read in the three bands, process them, and then display the transformed RGB image. This image can be saved in the standard PANCROMA formats as usual.
The standard image processing settings should work adequately in most cases. However, PANCROMA offers image processing utilities to improve the outcome if this is needed. To access these utilities, check the 'Activate Image Processing Routines' check box on the main menu after selecting your three band files. When the Image Processing Data Input Box appears, check the two '742 Band Adjust' check boxes.
The Red Cutoff, Blue Cutoff and Red Adjust track bars appear in the left data group. This set addresses the inherent lack of red tone in the non-vegetated areas. PANCROMA will increase the red DN value by the amount indicated on the Red Adjust track bar. This amount will be added only to pixels with red DNs above the Red Cutoff value and below the Blue Cutoff value.
The second adjustment in the left data group is the Ocean Cutoff track bar. PANCROMA will zero the red and green DN values for all pixels with a red DN less than the Ocean Cutoff value. This serves to mask the ocean areas that would otherwise have a green tone if these pixels were not thus processed.
The right data group simply has Red, Green and Blue adjust track bars. The DNs of these channels will be incremented or decremented according to the values on the track bars. The default settings are 8, 8, and 0 for Red, Green and Blue, respectively. A processed section of the above GEOCOVER file is shown below.
GEOCOVER can offer a very attractive option to standard pan sharpened Landsat scenes, and can provide an ideal data set for preparing maps or earth mosaics.