cere 
DIGITAL MAP REVISION IN A HYBRID GEOGRAPHIC 
INFORMATION SYSTEM 
Eugene Derenyi and Chee Hua Teng 
Department of Surveying Engineering 
University of New Brunswick 
Fredericton, N.B. Canada 
COMMISSION IV 
ABSTRACT 
It is advantageous to perform the updating of digital maps directly in a geographic information system (GIS). Through the research 
and development efforts at the University of New Brunswick, the Computer Aided Resource Information System (CARIS) has been 
equipped with an integrated digital image/map display and with orthoimage generation capabilities. Thus, map revision can be 
performed by on-screen digitization. Digital elevation models or piecewise rectification are used to correct for the relief 
displacement. Three map revision schemes were devised and then tested on 1:50 000 and 1:10 000 scale maps. 
KEY WORDS: Map revision, Orthoimage, GIS, Rectification DEM, Registration, Integrated system. 
1. INTRODUCTION 
In the 1980s, imagery acquired by photographic and 
electronic sensors was utilized in a distinctly different 
manner. Photographic camera products were handled in 
hard copy form. Quantitative analysis was performed in 
various photogrammetric restitution instruments, where 
graphical and digital map manuscripts or orthophotos were 
produced. Qualitative information extraction was performed 
by visual interpretation. Data acquired by non-photographic 
sensors were mostly handled in digital form and evaluated in 
stand-alone digital image analysis systems. The results were 
then recorded on hard copy outputs. This decade was also 
marked by the proliferation of geographic information 
systems (GISs). Most of them were established by 
digitization of existing hard copy maps. 
By the turn of this decade, GISs were widespread and firmly 
established; the conversion of hard copy images into digital 
form has become affordable and digital, soft-copy 
photogrammetry began to evolve as a viable alternative to 
analogue restitution instruments. It is imperative that the 
revision and updating process of maps and resource 
inventories stored in GISs reflects these new developments. 
The updating of a GIS can be accomplished ineither off-line 
or on-line mode. Off-line means a data capture in a device, 
such as a stereoplotter, digital image analysis system or an 
optical transfer scope, which is not an integral part of a GIS 
workstation. On-line means that the revision takes place in 
the GIS workstation itself, using digital images. This 
approach requires a hybrid GIS, which has both vector and 
raster data handling capability. The Computer Aided 
Resource Information System with Raster Image Extension 
(CARIS/RIX) is such a system. 
CARIS is a GIS marketed by Universal Systems Ltd. of 
Fredericton, N.B., Canada, while RIX was developed at the 
University of New Brunswick [Derenyi, 1991]. RIX 
supports the superimposition of vector data on a raster image 
backdrop, on-screen digitization, image classification, 
analytical photogrammetric operations and a variety of 
image processing and geometric registration tasks in 
interactive and batch mode. 
$33 
Three map revision schemes were developed and tested in 
this environment: 
* piecewise rectification, 
* DEM corrected tracing, and 
* digital orthoimage tracing. 
All three schemes are applicable to monoscopic images. 
2. MAP REVISION SCHEMES 
2.1 Piecewise Rectification 
This scheme is based on the assumption that all inherent 
distortions including the relief displacement of an image, can 
be reduced to an acceptable level by a two-dimensional 
transformation. This assumption is only valied in 
subregions of an image, where the magnitude of the relief 
displacement is below the tolerance set by the map accuracy 
standards. It is also assumed that the image and the map 
files are in a reasonably good registration. 
The residual mis-registration errors evident in subregions of 
an image, caused by deficiencies in the data used for the 
registration and by the neglected topographic effect, is then 
corrected by localized incremental transformation. In this 
process the analyst interactively improves the map to image 
registration by matching well-defined points and features 
within a small segment of the data surrounding the area of 
interest. It is done by incrementally translating, scaling, 
rotating and skewing the map (vector) to the image (raster). 
Although the transformation parameters are operated on 
sequentially, their effects are accumulated to form an affine 
transformation. Once a satisfactory registration has been 
achieved, the digital map file can be updated by digitization 
and editing in the image display. The inverse transformation 
is then applied to the coordinates of the new map objects, to 
register to the original map. 
This scheme is, in fact, the digital emulation of the technique 
used in Zoom Transfer Scopes. It is fast and simple. By 
temporarily warping the map to the image, the lengthy 
resampling operation of the raster data is avoided. This 
scheme is especially useful when changes in the planimetric 
content of the maps are concentrated in localized areas of the 
image.