with the support of Betsy Mackenzie, Oliver Weatherbee, Angela Watson, Larry Pomatto and Ken Richter
The State of Delaware contracted with PhotoScience Inc. to develop 1-meter resolution color-infrared digital ortho-photos for the entire state. Airphotos were taken in March, 1992, digitized and georeferenced to Delaware State Plane coordinates (meters, North American Datum 1983, based on GRS 1980 spheroid). The source data are 140 MB 24-bit color files for each of the 172 quarter-quads in Delaware. In these images the infrared reflectance (typically from vegetation) is displayed as red; red reflectance is displayed as green; green reflectance is displayed as blue.
To facilitate dissemination of these data, we resampled the 172 DOQ's covering the state to UTM coordinates (meters, NAD 1983) at 5-meter resolution in 8-bit color, achieving initial file size reductions of almost 99 percent. While the original DOQ series requires approximately 25 gigabytes of storage, our series totals approximately 310 megabytes, and is storable on a single CD in .BIP (SPOT Image) format (see below). We used a nearest-neighbor (i.e. center-cell selection) resampling procecure which generates a high-contrast image with some blockiness in very small features. Each transformation used four 3.75-minute-interval reference points located near the corners of each image. Each resampled image is approximately 1210 x 1510 pixels, with a 10-pixel (50-meter) border cropped out to eliminate slight edge skewness resulting from the reprojection. Images still have at least 500 meters of edge overlap.
Color palettes were brightened and adjusted for improved consistency using Adobe Photoshop. Sun glare has been masked out of water pixels in most coastal images in order to correct color palettes.
The resampled DOQ's are distributed from this page as high-quality JPEG-format images. Average image size is about 600K for each quarter quad (roughly 60% less than equivalent GIF files). While we believe we have achieved a near-optimal file compression (at a cost of slight lossiness) for efficient network delivery, these images may still take a few minutes to download, particularly if you are using a modem. Simply click inside any green-highlighted quarter quad in the reference map to view (download) an image. Or you can select from the alphabetical list of images.
Note that these images are larger than most monitors, so you may have to pan around a downloaded image using the horizontal and vertical display bars on edges of the image frame. To save a displayed image to your disk, click your right mouse button on it, select "Save Image As" and give the image any name you like. Netscape Navigator will save the displayed image in JPEG format; Internet Explorer will save it in Windows Bitmap (BMP) format. Saved as BMP's, these will be very large files with no gain in image quality. To download and save as JPEG with Internet Explorer, right-click on the link in the alphabetical list, not displayed image itself.
Here are suggestions for importing these data to various GIS's:
5.000000 --pixel size (meters) in X-dimension 0.000000 --row rotation factor 0.000000 --column rotation factor -5.000000 --pixel size (meters) in Y-dimension 435402.500000 --UTM easting coordinate of center of upper-left pixel 4393647.500000 --UTM northing coordinate of center of upper-left pixelSince these images use different color tables, they will not display correctly if loaded from an ArcView image catalog.
If your GIS supports image color-separation, you can create rough Normalized Difference Vegetation Index (NDVI) maps by ratioing the infrared and red reflectances: (IR-R)/(IR+R). NDVI's index vegetative biomass density. Check out the example (Fairmount SW quarter quad).
We would appreciate your comments or suggestions!
SPATLAB home page