INTEGRATION OF DIGITAL PHOTOGRAMMETRY AND RASTER GIS 
Kurt Novak 
Ohio State University 
Department of Geodetic Science and Surveying 
1958 Neil Avenue 
Columbus, Ohio 43210-1247 
614/292-6753 
Stephen L. Sperry 
ERDAS, Inc. 
2801 Buford Highway, Suite 300 
Atlanta, Georgia 30329-2137 USA 
404/248-9000 
Fax: 404/248-9400 
ISPRS Commission IV 
ABSTRACT 
This paper investigates the full integration of digital photogrammetric functions in a geographic information system (GIS). 
Photogrammetry is a very powerful tool that can be used to create accurate data for a GIS. The major obstacle against a 
wider application as a data capture method is the complicated instrumentation necessary. However, an implementation of 
photogrammetric mapping and triangulation functions on a workstation in an image analysis environment eliminates the cost 
  
  
of analytical instruments. This makes the system easy to use for non-photogrammetrists and it can be done directly in a data 
base environment. 
The concept of a digital photogrammetric plotter implemented as a software module of an existing GIS is presented. A full 
range of photogrammetric functions were developed in a workstation environment. They can be used to generate digital 
elevation models (DEMs) and digital orthophotos from aerial 
layers of a GIS. The developed software is not limited to di 
satellite triangulation. Stereo images are displayed on a works 
photographs and satellite imagery. These are two important 
gital mapping, but can be used for highly accurate aerial or 
tation screen where points can be selected semi-automatically. 
This allows analysts to accurately integrate satellite data with digitized aerial photographs in a common system. 
INTRODUCTION 
Although the geographic information system (GIS) has 
become a popular tool for many organizations, their 
implementation is often stalled by the lack of available digital 
data. Most GIS's are based on vectors (line data) digitized 
from existing maps. The digitized maps often include error— 
those transposed from incorrect or out-of-date maps, as well 
as those occurring during digitizing. Two critical GIS data 
layers, terrain and land cover, are typically difficult to develop 
in vector format. Contour maps representing terrain leave 
many areas blank. Land cover maps generated through 
manual air photo interpretation methods are time consuming 
and expensive. These data gaps can be filled by information 
from a raster GIS. 
Information for a raster GIS is mostly derived from satellite 
sensors or scanned aerial photographs which record spectral 
information about the surface of the earth. The raster 
equivalents of vector contours and land cover maps are digital 
elevation models (DEMs) and digital orthophotos. Digital 
orthophotos are image-based maps which have been corrected 
for relief displacement. 
In the past, DEMs and orthophotos were derived by manual 
photogrammetric techniques. Through the development and 
enhancement of digital photogrammetry, it is possible to 
derive the information fully automatically from digital 
imagery. Because of the similarity in data formats between 
digital photogrammetry and GIS, the DEM and digital 
orthophoto can be readily used in raster GIS's. Due to recent 
advances in the integration of raster and vector GIS, the data 
are also available to vector GIS's. 
894 
Digital photogrammetry provides not only current and accurate 
information for GIS's, but its algorithms are also useful for 
extracting products from existing GIS data layers. These 
products include perspective views, slope models, exposition 
models and contours. This paper describes the development 
of a digital photogrammetric module in a raster GIS. First, the 
functions by which digital photogrammetry can enhance the 
capabilities of automatic information extraction are discussed. 
This is followed by a discussion of a specific implementation 
of photogrammetric algorithms in a GIS. The discussion 
includes descriptions of sensor orientation, image matching, 
DEM interpolation and digital orthophoto generation. Typical 
problems that are encountered in these processes are identified 
along with operator solutions. In the conclusions, future 
developments and further enhancements of digital 
photogrammetry are discussed in terms of new 
photogrammetric image analysis techniques. 
DIGITAL PHOTOGRAMMETRY 
Photogrammetry has been used for about one hundred years to 
generate maps from aerial photographs and to compute highly 
accurate point positions in three dimensions. It is a matured 
technique for extracting spatial data. The advance of digital 
sensors have made it possible to use photogrammetry on a 
computer and apply its algorithms to digital imagery. The use 
of a computer in photogrammetry is known as a softcopy 
photogrammetry system. Currently, the most popular 
techniques are image matching and DÉM generation, as well 
as digital orthophotography.