Aerial photographs represent a large and heavily used portion of the Map & Imagery Laboratory's spatial data collection. In order to provide these data through the ADL catalog, they must be digitized and their spatial extent must be determined and defined as geographic coordinates for inclusion in their metadata. Aerial photographs most commonly exist as positive prints, positive transparencies, and negative transparencies. Film types are most commonly black and white, color and color infrared, and size is predominantly 9" x 9".
The scanning station is a Pentium-level PC and runs a Sharp JX-610 large format flatbed scanner with optional transparency attachment through Adobe PhotoShop. In a trade-off between image resolution and file size, we have been scanning the aerial photographs at 600 pixels per inch. Files are saved in TIFF format without compression to the local hard drive until approximately 1 Gigabyte of scanned files have been generated (approximately 30 gray-scale images or 10 full color images). The files are then written to 8mm tape for archiving. File naming convention preserves aerial photograph frame name for easy tracking.
The scanning operation is a documented procedure that can be executed by low-level workers such as students.
Spatial coordinate determination and definition procedures vary. A quick and dirty procedure can be employed when aerial photograph indexes exist in which the frame boundaries and orientation are readily discernable. This procedure involves tablet digitization through Arc/Info. The aerial photograph index is registered to the tablet using known ground control points and the frame corners are then digitized for the entire index. The resultant coverage is then projected to geographic coordinate space and finally ``ungenerated'' to yield ASCII pairs of latitude and longitude in decimal degrees. These coordinates are then ready to be incorporated into the aerial photograph's metadata. Without the presence of an appropriate index, however, coordinate determination and definition must be done individually for each frame. Currently, students with GIS software experience are performing these procedures.