PHOTOGRAMMETRIC ENGINEERING & REMOTE SENSING
February 2016
87
Q: What are the situations that require us to redo the
boresighting of the aerial camera/inertial measurement
unit (IMU) misalignment angels? We understand that a
new boresight determination is needed, only if, the IMU
is removed from the camera and re-attached again.
However, we recently encountered a group of experts on
the subject who suggested doing boresight calibration
for every mission in the project area to increase the
accuracy. This poses a great burden to our operations
and budget. If we have to fly eight different blocks with
eight different missions, wewill have to do eight boresight
calibrations. Does it really increase accuracy if we
boresight the IMU for every mission?
Mulualem Yeshitila, Ethiopia
Dr. Abdullah:
Before I address the issues raised in the question,
I would like to elaborate on IMU and the need for its boresight
calibration for readers who are not familiar with the subject. The
IMU, which was introduced to the mapping industry in the late
90s, is a system that uses integrated GPS and inertial technology
to measure camera attitude to accuracy sufficient to a certain
degree to support photogrammetric mapping. The IMUmeasures
three axes rotational angles in space for object that is rigidly
attached to it. The IMU senses gravity andmeasures acceleration
using sets of gyros and accelerometers. Therefore, it has its own
reference frame that is different from the camera body frame,
the mapping frame, and the mapping coordinates system. For
the rotational angles measured by the IMU to be translated to
rotation angles referenced in the camera frame and subsequently
in the mapping frame, the rotational angles between the camera
body frame and the inertial body frame need to be determined
accurately. The process of determining such rotation is what we
refer to as “boresighting”, “IMU misalignment calibration”, or
“IMU boresighting”. In addition, the GPS unit and the IMU are
physically separate from the perspective center of the camera
lens, and a constant displacement exists between their positions.
The extent of this displacement also needs to be determined using
conventional surveying techniques prior to the flight missions.
The latter displacement is usually referred to as the “lever
arm.” On the other hand, the rotational IMU misalignment is
determined by performing an aerial triangulation process on a
small controlled block of imagery usually referred to as the “bore-
sight area.” The exterior orientation parameters computed from
the aerial triangulation of the boresight area are then used by
the IMU processing software to determine the offset values and
export the rotational angles (omega, phi, kappa) for every image
of the block according the photogrammetric or mapping frame.
As for the question on whether an IMU boresighting is needed
with every block or mission, my direct answer is no you do
not always need to. The importance of the IMU boresighting
quality and frequency depend on how the IMU data is used in
the mapping process. IMU can be used in two ways during a
mapping process. Those are:
1. Direct Sensor Orientation: Here, the IMU-derived exterior
orientation angles (omega, phi, kappa) of the camera are
fed directly into the photogrammetric plotter or software to
set up the stereo models or to perform ortho-rectification.
In other words, there is no aerial triangulation in the
workflow. In this scenario, the quality and the accuracy
of the boresight angles are very critical for obtaining
parallax-free stereo geometry. The need and the frequency
of the IMU boresighting are widely disputable as IMU
manufacturers believe that their systems are stable
enough and do not require a boresight determination
with every mission. However, users’ experiences defeated
manufacturers’ claims. In many cases, users found
boresight determination was needed more frequently than
the manufacturers claimed. I do not believe that too many
mapping businesses—if any—produce mapping products
“The importance of the IMU boresighting
quality and frequency depend on how
the IMU data is used in the mapping
process.”
Photogrammetric Engineering & Remote Sensing
Vol. 82, No. 2, February 2016, pp. 87–88.
0099-1112/16/87–88
© 2015 American Society for Photogrammetry
and Remote Sensing
doi: 10.14358/PERS.83.2.87
