PHOTOGRAMMETRIC ENGINEERING & REMOTE SENSING
February 2020
85
Chapter 5: Quality Assurance and Quality Control of Remote Sensing Systems
INTRODUCTION
In the last few years, the mapping community has witnessed significant
developments in using passive and remote sensing technologies
onboard space-borne, airborne, and terrestrial platforms to provide
a wide range of products. These developments can be ascribed to
1) proliferation of high resolution space borne imaging satellites
operating in different portions of the electromagnetic spec-trum,
2) reduced cost and improved performance of modern color,
multispectral, and hyperspectral digital cameras, 3) continuous
developments in Light Detection and Ranging (LiDAR) systems,
4) capability of integrated Global Navigation Satellite Systems/
Inertial Navigation Systems (GNSS/INS) in providing accurate
position and orientation information for the utilized platforms, 5)
incorporation of multiple cameras and/or laser scanners onboard a
single platform, 6) emergence of non-traditional mapping platforms
such as airborne and terrestrial unmanned autonomous systems
- UAS, 7) convergence of research efforts from the mapping
and computer vision communities, and 8) increased demand for
geospatial data to satisfy the needs of non-traditional applications
(e.g., precision farming, infrastructure monitoring, powerline
clearance evaluation, and construction en-gineering management).
Taking advantage of such developments in the remote sensing
technologies is only possible when standard Quality Assurance and
Quality Control (QA/QC) procedures are in place to ensure the ut-most
precision of the mapping product. In this chapter, the term “Quality
Assurance – QA” is used to denote pre-mission activities focusing on
ensuring that a process will provide the quality needed by the user.
On the other hand, the term “Quality Control – QC” is used to denote
post-mission procedures for evaluating the quality of the final product.
QA mainly deals with creating manage-ment controls including the
calibration, planning, implementation, and review of data collection
ac-tivities. For example, a LiDAR-based QA activity should entail
gaining prior knowledge of the area to be surveyed in terms of its
extent and terrain coverage (e.g., vegetation and buildings) to set up the
appropriate flight configuration and specifications. In forested areas,
a slower speed, smaller scan angle, higher overlap percentage, and/
or lower flying height might be necessary to increase the point density
and to have more pulses penetrating to the ground. Also, the selection
of the appro-priate mission time according to the GNSS satellite
constellation distribution is another important QA item. For example,
a typical requirement is to have at least four well-distributed satellites
with elevation angles above 15° throughout the survey. Moreover, it is
recommended that the aircraft should stay within a given distance from
the GNSS base station. Another QA activity for LiDAR mapping is
the system calibration. For such an activity, one should have access to
the original ob-servations (GNSS, IMU, and the laser measurements)
or at least the trajectory and time-tagged point cloud. Such quantities
might not be always available to the end user. This Chapter will be
mainly focusing on the system calibration component of QA activities.
For an illustration of standard QC activities, one can refer to the well-
established photogrammet-ric procedures for evaluating the internal/
relative and the external/absolute accuracy of the final product. For
the evaluation of the internal/relative quality (IQC) of the outcome
from a photo-grammetric reconstruction exercise, we typically use
the a-posteriori variance factor and the vari-ance-covariance matrix
resulting from the bundle adjustment procedure. As for the exter-
nal/absolute quality (EQC) evaluation, checkpoint analysis using
independently measured targets is usually performed. Since the
computation of the LiDAR point cloud is not based on redundant
measurements, which are manipulated in an a
djustment procedure,
standard photogrammetric IQC measures are not possible. Moreover,
the irregular and sparse nature of the LiDAR point cloud makes the
EQC process more challenging. A commonly used EQC procedure
compares the LiDAR surface with independently collected control
points. Besides being expensive, this procedure does not provide
accurate verification of the horizontal quality of the LiDAR points,
unless specifically designed targets are utilized. Such inability is a
major drawback since the horizontal accuracy of the LiDAR points
is known to be inferior to the accuracy of these points in the vertical
direction. In this regard, this Chapter addresses the validation of remote
sensing data from space borne, airborne, and terrestrial platforms.
The chapter is organized as follows:
•
Section 2: Geometric Calibration / Validation of High Resolution
Imaging Sensors deals with the operational principles of modern
high resolution imaging satellites as well as the geometric
calibration and validation procedures which are adopted by
the system manufac-turers and end users, respectively.
• Section 3: Geometric Calibration / Validation of Airborne
and Terrestrial Optical Imaging Systems provides the
operational principles of single-head and multi-head airborne
and ter-restrial optical imaging systems together with
their geometric calibration and validation in the presence
or absence of onboard direct geo-referencing units.
• Section 4: Radiometric Calibration of Passive Optical
Imaging Systems focuses on radio-metric calibration
and correction aspects of airborne and terrestrial, mobile
passive imaging systems which operate in the reflective
or thermal ranges of the electromagnetic spectrum.
• Section 5: Geometric Calibration for Active Remote Sensing
Systems (LiDAR provides im-portant aspects for the geometric
calibration of LiDAR systems while considering the possi-
bility that the raw measurements are not always available.
• Section 6: Geometric Quality Control of LiDAR Data deals
crash land (“belly landing”). The multi-rotor types have the
capability for vertical take-off and landing (VTOL), a feature of
particular interest as they can operate in tighter and in-door spaces,
around structures, and they have high maneuvering and hovering
abilities. There also are hybrid fixed-wing systems, VTOL UAV,
such as the SONGBIRD (Thamm et al., 2015). For propulsion,
electric (including solar) and gas /diesel engines are used.
