INVESTIGATION OF AERIAL TRIANGULATION AND SURFACE GENERATION USING A SOFTCOPY 
PHOTOGRAMMETRIC SYSTEM 
Professor Frank L. Scarpace and Raad A. Saleh 
Laboratory for Softcopy Photogrammetric Systems 
Environmental Remote Sensing Center 
University of Wisconsin-Madison 
1225 W. Dayton St., Madison, WI 53706, USA 
Tel. (608)262-1585, Fax (608)262-5964, E-mail: scarpace@macc.wisc.edu, raad@cae.wisc.edu 
InterCommission WG IUIII 
KEY WORDS: Softcopy, Systems, DEM/DTM, Aerial Triangulation, Surface Generation, Large Scale Mapping. 
ABSTRACT: 
The accuracy of softcopy aerial triangulation and surface generation was investigated. The experiments were imple- 
mented using a high-end, commercially available softcopy photogrammetric system. Automated and semi-automated 
methods of data collection were used for point measurements, and surface generation. The results from the softcopy 
approach were compared to those produced by conventional analytical plotters. 
The data set used for the experiment was a strip of 65 low altitude photography with a scale of scale of 1:3600. These 
photographs were scanned using two high-end photogrammetric scanners. Ground coordinates of 263 pug points were 
measured automatically and aerial triangulated using the softcopy systems. Surface elevation data were also automati- 
cally generated for a 5 photo subset of the strip. The surface data and triangulated pug points were compared to data 
measured manually and extracted using analytical plotters. 
Analysis of these comparisons shows that accuracy of aerotriangulated points is better than 0.1 meters with a standard 
deviation of the differences between 0.1 and 0.2 meters. The average discrepancy between the two approaches in sur- 
face collection is close to zero, but the standard deviation in the automatically derived elevations are about 0.75 meter. 
This study concludes that, for large scale mapping, aerial triangulation results from softcopy systems are as accurate as 
those expected from conventional methods. This study also finds that the automatically generated elevation data are 
less accurate than that manually collected with an analytical plotter. 
1. INTRODUCTION cial softcopy photogrammetric system. The system in- 
vestigated is the Leica Helava DPW770. 
A research project entitled “Research and Development 
in Softcopy Photogrammetry" has been underway for 
two years at the Laboratory for Softcopy Photogrammet- 2. PROJECT DESCRIPTION 
ric Systems of the University of Wisconsin-Madison. 
The photographic coverage and ground controls used in 
This research has been funded by the Wisconsin De- this research were provided by the Technical Services 
partment of Transportation (WisDOT). The primary Office of WisDOT. WisDOT provided AT results and a 
objective is directed at investigating critical aspects of digital terrain model (DTM). Both of these products 
softcopy photogrammetry and presenting recommenda- were generated with analytical plotters using manual 
tions to WisDOT as to develop strategies for conversion methods. The ground coordinates of 263 pug points 
to softcopy technology. Aspects of this research in- along the strip were calculated using conventional AT. 
volved a review of the state of softcopy technology, a The DTM was collected using a Kern DSR-14 analyti- 
survey of commercial photogrammetric systems, and a cal plotter for a 5 photo subset of the strip. 
survey of photogrammetric scanners. 
The photographic coverage consisted of 65 photographs 
This paper describes experiments undertaken during the distributed over six strips, overlapping at the end of 
project to investigate the accuracy of aerial triangulation each strip, thus forming a snake-like block. The 65 
(AT) and surface generation implemented in a commer- diapositives were scanned with the Helava DSW200 at a 
340 
International Archives of Photogrammetry and Remote Sensing. Vol. XXXI, Part B2. Vienna 1996 
eh ppd REM — M D 
"Cj A 
Ta 
grec 
ver 
A 
RN