A New Pseudorigorous Lidar System
Calibration Strategy with Minimal
Requirements and Higher Capability
M. Ritelli a
Abstract
Most times, rigorous airborne Light Detection and Ranging
(LiDAR) system calibration is not possible because the raw
measurements are unavailable, and pseudorigorous app-
roaches that synthesize the raw measurements from the
point cloud (and in some cases the trajectory) are used. The
Quasi-Rigorous/Quasi-Simplified
is a new pseudorigorous
approach proposed here which is more useful when com-
pared to the existing pseudorigorous approaches (
Simpli-
fied
and
Quasi-Rigorous
) because it has the same minimal
data requirements as the
Simplified
but provides the higher
capabilities of the
Quasi-Rigorous
. The experimental results
compare the
Quasi-Rigorous/Quasi-Simplified
approach to
the existing pseudorigorous (
Simplified and Quasi-Rigorous
)
and rigorous approaches. When compared to the exist-
ing approach with minimal requirements
(
the
Simplified)
,
the results show that the
Quasi-Rigorous/Quasi-Simplified
approach is equally successful in significantly reducing
the impact of systematic errors and offers additional cap-
abilities in that it can be used on any type of terrain, can
use nonparallel flight lines, and can incorporate control.
Introduction
A Light Detection and Ranging (
LiDAR
) system refers to the
laser-ranging and the integrated direct geo-referencing units.
The direct geo-referencing unit is comprised of a Global Navi-
gation Satellite System (
GNSS
) and an Inertial Measurement
Unit (
IMU
). In addition to each component of the
LiDAR
system
having many advances in the past decade, airborne platforms
have had significant technological growth (e.g. Unmanned
Aerial Systems). Together, these advances have accelerated
the growing use and application of
LiDAR
systems because
they allow for higher accuracy, and the systems are more
readily available at a lower cost. With the large user commu-
nity and its continual increase, development of standard op-
erating procedures for
LiDAR
system modeling and calibration
will ensure that the systems consistently meet industry stan-
dards (to the best of the authors’ knowledge, these standards
are still under public review and can be referenced at
LIDAR
Standards and Guidelines Complete Public Review Process,
2017). To achieve this, rigorous mathematical models are used
which relate all system parameters and system measurements
from each component to calculate each point’s ground coordi-
nate. The appropriate model should also include parameters
that account for systematic errors within the
LiDAR
system.
In addition to system modelling, quality assurance (
QA
) and
quality control procedures are needed to maintain quality
throughout all planning, collection, and data delivery steps
(Habib
et al.
2009a). A prominent
QA
procedure is the in-flight
rately decouple
ing the system
parameters, the calibration process minimizes the impact of
systematic errors on the resulting point cloud.
The existing approaches to eliminate and/or reduce the
effect of systematic errors are categorized as either system-
driven (calibration) or data-driven (strip adjustment). System-
driven approaches are superior because they constrain point
cloud reconstruction to the geometric relationship (i.e. the
sensor model) that exists between the
GNSS
/inertial navigation
system (
INS
), the
LiDAR
scanning mechanism, and the
LiDAR
footprint on the ground. While system-driven approaches
preserve the link to the sensor model, data-driven approaches
arbitrarily fit data strips together through a transformation
model that may not scale appropriately to the entire dataset.
The data-driven approaches for eliminating discrepancies are
never a suitable substitute to using system-driven procedures,
but they are sometimes employed because the end user does
not have access to the raw measurements (the raw measure-
ments being referred to are the integrated
GNSS
/
INS
position
and orientation information, and the
LiDAR
unit measure-
ments). Typically, the data provider has sole access to the raw
measurements and their
LiDAR
system calibration is consid-
ered a trade secret. Overall, there is a need for standardized
LiDAR
system calibration procedures which are system-driven,
even in the absence of raw measurements, and are also gener-
al for the wide range of users. The system-driven approaches
can be further categorized as
rigorous
or
pseudorigorous
. The
rigorous
approaches are used when all raw measurements are
available, and the
pseudorigorous
approaches are used when
there is a full or partial raw measurement unavailability.
There have been various types of system-driven calibra-
tion approaches developed that differ based on the need for
control surfaces (as opposed to using overlapping strips),
dependence on urban settings, manual operations, and/or
preprocessing of data to extract certain linear/planar features.
If control surfaces were readily available and economical,
approaches that constrain the
LiDAR
point to those control
surfaces (Filin 2001) would be ideal, but this is not always the
case and calibration solutions that take this approach are not
suitable for all users. The majority of calibration methodol-
ogy development has been done on the premise of comparing
overlapping strips (to avoid dependence on control surfaces)
for both r
igorous
approaches (Skaloud and Lichti 2006; Friess
2006; Kersting 2011; Kersting
et al.
2012) and
pseudorigor-
ous
approaches (e.g.
Simplified
and
Quasi-Rigorous
) (Burman
2002; Toth 2002; Morin 2002; Habib
et al.
2010; Bang 2010).
M. Ritelli and A. Habib are with the Digital Photogrammetry
Research Group, Lyles School of Civil Engineering, Purdue
University, 550 Stadium Mall Dr., West Lafayette, IN 47907
(
;
).
Photogrammetric Engineering & Remote Sensing
Vol. 85, No. 4, April 2019, pp. 313–321.
0099-1112/18/313–321
© 2019 American Society for Photogrammetry
and Remote Sensing
doi: 10.14358/PERS.85.4.313
PHOTOGRAMMETRIC ENGINEERING & REMOTE SENSING
April 2019
313
