Page 228 - Digital Analysis of Remotely Sensed Imagery

192 Cha pte r F i v e

5.8 Image Direct Georeferencing
With the advent of GPS technology, it is possible to couple it with an
inertial navigation system (INS), also known as an independent
measurement unit (IMU), in acquisition of remote sensing data. In
addition to easing aircraft navigation, this integration considerably
facilitates georeferencing of remote sensing imagery. For instance,
during flight to acquire aerial photographs there is no need for the
pilot to follow a rigid position thanks to “in-flight alignment” afforded
by GPS. The INS is able to provide a continuous high-bandwidth
measurement of position and velocity after the noisy velocity from
GPS outputs is smoothed (Skaloud, 2002). The generated information
on sensor position and exterior orientation at the time of imaging
from the deployment of a GPS-aided INS makes it possible to directly
georeference images without ground control (Schwarz et al., 1993).
Image direct georeferencing is a process of restoring the image orienta-
tion from in-flight measured exterior orientation parameters of the
sensor without reliance on ground control. The position of all pixels
on this restored image can be translated into ground coordinates
according to their internal mathematical relationship.
The concept behind GPS-aided INS for direct georeferencing
dates back to the late 1980s and early 1990s, with the first system
commercialized in 1996. Since then tremendous progress has been
made, with the capabilities of this new technology fully exploited.
Now, image direct georeferencing has been accepted as an augmen-
tation to and replacement of aerial triangulation. With advances in
computing and the wide use of digital cameras in aerial photogra-
phy, direct image georeferencing is quickly becoming the de facto
industry standard. It has evolved to such a degree that the tradi-
tional workflow of data acquisition, data processing, and map pro-
duction can be accomplished in one step, thus revolutionizing our
perspective of mapping science altogether. High quality mapping
products are generated in a much simplified process. In this section
the principle of direct georeferencing is presented first, followed by
a comparison of its performance with the conventional GCP-based
image rectification.

5.8.1 Transformation Equation
As illustrated in Fig. 5.20, the basic concept of direct georeferencing is
expressed mathematically as

R = R + s R R INS r (5.28)
A O a INS c
where R is the coordinates of point A, (E N H ) or its georefer-
T
A A A A
enced position in the ground coordinate system, and R is the 3D
O
T
coordinates (E N H ) of the exposure center of the imaging sensor
O O O