tical
the
nage-
ffina-
3 of
dels.
dels
lo-
lation
hods:
ems,
the
for
the
stem
s in
nul-
is
] of
her
sion
°an-
nts
aps,
imu-
elli-
ter
ige-
IRR
àS—
up
re-
de-
ass
and
les.
r
APPLICATION OF DIGITAL CAMERAS AND
GPS FOR AERIAL PHOTOGRAMMETRIC
MAPPING
Kurt Novak
Department of Geodetic Science and Surveying, Center for Mapping
The Ohio State University
Commission I
ABSTRACT
This paper describes the development and application of a fully digital, aerial image
acquisition system which is integrated with GPS. A high resolution digital camera captures
overlapping vertical images from an airplane. The exposure stations are tagged with the
GPS time and the accurate position of the airplane. During post-processing images and
navigation data are integrated. All information is stored in an image data-base related to a
GIS. Kinematics GPS is applied to determine the exposure stations within ten centimeters.
An aerial triangulation solves for additional camera parameters and is controlled by the GPS
locations of the perspective centers. The major application of this new system will be the
mapping of utility lines, roads, pipelines, and the generation of digital elevation models and
orthophotos for engineering applications.
Keywords: Aerotriangulation, Camera, Digital Systems, GPS, Integrated System
1. INTRODUCTION
The development of integrated mobile mapping systems
is a major research focus at the Center for Mapping of the
Ohio State University. The most famous so far is probably
the GPS-Van, which is a vehicle-based mapping system that
combines GPS, inertial sensors and a digital stereo-vision
system for creating highway inventories (Novak, 1991;
Bossler et al., 1991). The same sensors used in the GPS-
Van can also be implemented in an airplane. Some hardware
components are not absolutely necessary in the airplane,
such as the inertial system or a second camera, as the
parameters measured by these devices can be easily
recovered by analytical triangulation techniques.
The digital aerial mapping system described in this paper
was named MapCam. It consists of a fully digital, high-
resolution, frame CCD camera that can capture imagery at
pre-defined times and store the data on a digital tape. A GPS
receiver operating in kinematic mode allows to tag the
images with the accurate position of the airplane; it is applied
during post-processing to control aerial triangulation. All
information captured during the flight is stored in a GIS;
flight-lines and exposure stations define a geographic
reference, and the images are stored relative to these
locations as attributes. Any image can be displayed on the
computer-screen by selecting its geographic location on a
digital map. All photogrammetric operations, such as image
coordinate measurement, aerial triangulation, DEM and
feature extraction can be done on a post-processing
workstation semi-automatically. The extracted information
is directly available in the GIS.
GPS controlled aerial triangulation is being applied
operationally by various organizations (Lapine, 1990). The
purpose of GPS is to eliminate ground control for aerial
mapping; to our knowledge, however, all applications relate
to metric, aerial, film cameras, so that the analysis of the
imagery must be done by an operator on an analog or
analytical stereo-plotter.
Electronic cameras have been applied for remote sensing
purposes in airplanes. In most cases low resolution, analog
video cameras were used. Some special sensors have been
developed for digital, aerial mapping, such as the MOMS
(Ebner et al., 1991) camera. Both are based on pushbroom
type CCD arrays, which provide in-flight stereo by a
vertical, as well as a forward and an aft-looking scan. The
high price of these systems, which are mostly at the
prototype stage, and the complicated geometrical camera
model prohibited their wide distribution to date. To our
knowledge nobody experimented with high-resolution frame
CCD cameras together with GPS in airplanes.
In this paper various aspects of the development of
MapCam are discussed. In the next chapter the hardware
components are described. Then we explain the
mathematical model which we applied to perform aerial
triangulation and camera calibration. Some products, such
as DEMs and digital orthophotos, are derived and displayed
in a 3-dimensional perspective view. In the final chapter
potential applications are shown to demonstrate the
versatility of MapCam.
2. HARDWARE COMPONENTS
The MapCam system consists of three major
components: a digital, high resolution CCD camera, a GPS
receiver and a computer-control and storage unit. Our CCD
camera is a Kodak Hawkeye M-3 (figure 1), which
incorporates a solid state CCD sensor of 1280 (H) x 1024
(V) pixels in the body of a regular Nikon F-3 camera. The
exposure is controlled by the electronics of the Nikon
camera. It applies a slit-type focal plane shutter. The CCD
sensor transfers digital data to a frame buffer, which is
installed in a separate box together with a portable harddisk
that holds up to 120 images. The Hawkeye M-3 camera can
be operated from a battery and can be easily carried around.
In order to circumvent the limitations of the harddisk we
connected the data-capture box to a digital tape drive
(Exabyte) through a SCSI interface. We also included a
data-compression board to reduce transfer rates and save
storage space. One Exabyte tape holds up to 5 GBytes of
uncompressed data which corresponds to 3,850 images.
Currently data-transfer is limited to one image per second.
We experimented with two different types of GPS
systems: for metric mapping applications we used a pair of
Trimble 4000 ST survey-quality receivers. They operate in
kinematic mode, which means that one is positioned over a
known base-station, the other is mounted on the fuselage of
our top-wing airplane. They are both observing phases of
the GPS carrier signal, and provide a clock for
synchronizing all components of MapCam. With these type
of receivers we can completely eliminate ground control for
aero-triangulation, as the exposure stations can be
determined to better than 10 cm. However, satellite lock
must be maintained continuously once the airplane's GPS
antenna was initialized over a known target on the runway.
This means that the pilot must fly very wide turns without
banking the airplane.
