u 2004 
pefully 
ated to 
new 
ized. 
or their 
pertise. 
r their 
flights. 
/hat Is 
1i. 
th 
EEPE 
ind 
#59, 
ung 
sium 
th 
9-12 
resight 
ation, 
mber. 
2610, 
Ma 
d. 
PERFORMANCE ANALYSIS OF INTEGRATED SENSOR ORIENTATION 
A. W. L. 1p**, N. El-Sheimy*, J. Hutton " 
“ Mobile Multi-Sensor Research Group, Department of Geomatics Engineering, University of Calgary, 2500 University Drive NW, 
Calgary, Alberta, T2N 1N4 Canada — awlip@ucalgary.ca, naser@geomatics.ucalgary.ca 
? Applanix Corporation, 85 Leek Crescent, Richmond Hill, Ontario, L4B 3B3, Canada — jhutton@applanix.com 
TS SS 3 
KEY WORDS: Aerial Mapping, Aerial Triangulation, Direct Georeferencing, GPS/INS 
ABSTRACT: 
Integrated multi-sensor systems, with their major progress in terms of sensor resolution, data rate and operational flexibility, have 
become a very attractive mapping tool over the last decade. In the aerial mapping application, for example, exterior orientation 
parameters (EO) for the imaging sensors are required. Using Differential Global Positioning System (DGPS) with Inertial 
Measurement Units (IMU), direct determination of the EO parameters can be obtained from the inertial/GPS navigation solution. 
This process is referred to as Direct Georeferencing (DG). Direct Georeferencing provides substantial benefits over the indirect 
determination method of estimating EO parameters from conventional aerial triangulation (AT) techniques using block of images 
with sufficient number of known control points. These benefits include the ability to map remote and inaccessible regions, and by 
replacing tie point measurements/matching and AT, significant cost-savings can be obtained for projects that do not require stereo 
models (such as projects with existing DEM, single image or strip/corridor mapping). The accuracy of Direct Georeferencing 
however, is limited by the accuracy attainable by the DGPS, IMU and any residual datum calibration errors. These can typically be 
as large as 10 cm RMSE, which is not sufficient for some large scale mapping applications. However, by combining the direct EO 
data in a traditional block adjustment, AT techniques can be used to remove the residual errors in the solution. This technique is 
known as Integrated Sensor Orientation (ISO). It has several advantages over traditional AT, primarily since; the stable geometry 
provided by direct EO can reduce the number of required GCP and tie-point to a minimum. At the same time, ISO provides an 
excellent means to QA/QC the EO from a DG system. 
This paper will examine the factors that determine the system performance for ISO. In addition, an example will be given to 
illustrate the expected accuracy of an aerial mapping project using ISO under different qualities of DGPS/IMU data. 
I. INTRODUCTION 
During the recent revolution in aerial mapping, two major 
components have undergone rapid research and development. 
The first is the development of digital imaging sensor and the 
second is the determination of exterior orientation parameters 
(EO) using integrated navigation systems. Digital imaging 
sensors considerably reduce the data processing effort by 
eliminating the digitizing step. They also open the way towards 
new and flexible designs of the processing chain, making ample 
use of mathematical software tools readily available. In the 
form of digital frame cameras, they are inexpensive enough to 
make redundancy a major design tool. Precise integrated 
navigation systems, in the form of DGPS/IMU, has developed 
to a point where it can provide the solution of the exterior 
orientation problem without the use of ground control points 
(GPC) or block adjustment procedures. The also open the area 
for new mapping sensors, such as line scanners, SAR and 
LIDAR, where EO cannot be obtained using the traditional AT 
techniques. This paper will focus on the second component; the 
determination of EO parameters for full frame digital imagining 
sensors. 
There are two basic approaches for computing the EO of an 
imaging sensor, they are: 
  
* Corresponding author. 
* Determine the EO directly using suitable position and 
orientation sensors 
* Determine the EO in-directly by extracting them from 
a block of images with a sufficient number of known 
ground control points. 
The fist approach is known as Direct Georeferencing, most 
commonly achieved using an integrated DGPS/Inertial system. 
Such systems have been well studied and implemented 
commercially, such as Applanix's Position Orientation System 
for Airborne Vehicles (POS AV). The second approach uses AT. 
which relies on a network of tie points in a block of frame 
imagery with a sufficient number of known ground control 
points. 
When the availability of ground control points is in question, 
such as in forest and desert areas or along a coastline, the ability 
of resolving the EO parameters in-directly is limited. Often 
these areas are also very important when an emergency 
response has to be taken. Such application requires fast 
orthophoto generation, and there is insufficient time and 
resource to extract EO parameters using traditional AT. In 
addition, some projects only require a single strip or single 
photo orientation, such as in the case where there is an existing 
Digital Elevation Model or DEM, Here the use of traditional 
AT to determine EO parameters is unpractical because it