203 
ERROR BUDGET OF LIDAR SYSTEMS AND QUALITY CONTROL OF THE DERIVED 
POINT CLOUD 
A. F. Habib a ’ *, M. Al-Durgham a , A. P. Kersting 1 2 , P. Quackenbush b 
a Dept, of Geomatics Engineering, University of Calgary, 2500 University Dr. NW, 
Calgary, Alberta, T2N 1N4, Canada, habib@geomatics.ucalgary.ca 
b Base Mapping and Geomatic Services (BMGS), 2 nd Fir. 395 Waterfront Crescent, Victoria, BC, Canada, V8T 5K7, 
Paul. Quackenbush@go v. be. ca 
Commission, WG 1/1 
KEY WORDS: LiDAR, sensor model, error analysis, accuracy, quality inspection. 
ABSTRACT: 
The improving capability of the direct geo-referencing technology is having a positive impact on the widespread adoption of LiDAR 
systems for the acquisition of dense and accurate surface models over extended areas. A typical LiDAR system consists of three 
main components: a GPS system to provide position information, an IMU unit for attitude determination, and a laser unit to provide 
the range (distance) between the laser-beam firing point and the ground point (laser footprint). The measured ranges are coupled 
with the position and attitude information from the GPS/IMU integration process as well as the bore-sighting parameters relating the 
system components to derive the ground coordinates of the LiDAR footprints. Unlike photogrammetric techniques, the derivation of 
the point cloud from the LiDAR measurements is not a transparent process. In other words, the raw system measurements are not 
always provided to the system user. Moreover, the coordinate computation of the LiDAR footprints is not based on redundant 
measurements, which are manipulated in an adjustment procedure. Consequently, one does not have the associated measures (e.g., 
variance component of unit weight and variance-covariance matrices of the derived parameters), which can be used to evaluate the 
quality of the final product. This paper is concerned with providing a tool for the quality control (QC) of the LiDAR point cloud. 
The objective of the QC procedure is to verify the accuracy of the LiDAR footprints. In other words, the QC procedure would test 
whether the expected accuracy has been achieved or not. The paper will start with a brief discussion of the LiDAR mathematical 
model relating the system measurements to the ground coordinates of the point cloud. Then, an analysis of possible systematic and 
random errors and their impact on the resulting surface will be outlined. Following the discussion of the error sources and their 
impact on the accuracy of the LiDAR footprints, a QC tool will be proposed. The paper will conclude by experimental results from a 
real dataset involving overlapping strips from operational LiDAR system. 
1. INTRODUCTION 
The direct acquisition of a high density and accurate 3D point 
cloud has made LiDAR systems the preferred technology for 
the generation of topographic data to satisfy the needs of 
several applications (e.g., digital surface model (DSM) creation, 
digital terrain model (DTM) generation, orthophoto production, 
3D city modeling, and forest mapping). A typical LiDAR 
system consists of three main components: a GPS system to 
provide position information, an IMU unit for attitude 
determination, and a laser unit to provide the range (distance) 
between the laser-beam firing point and the ground point (laser 
footprint). The measured ranges are coupled with the position 
and attitude information from the GPS/IMU integration process 
as well as the bore-sighting parameters relating the system 
components to derive the ground coordinates of the LiDAR 
footprints. Although the use of LiDAR data for different 
applications has increased significantly in the past few years, 
the user community still lacks standard and efficient procedures 
for evaluating the quality of the provided point cloud. 
Compared to photogrammetric and other surveying techniques, 
the computation of the LiDAR footprints is not based on 
redundant measurements, which are manipulated in an 
adjustment procedure. Consequently, standard measures for 
evaluating the quality of the final product, such as the a- 
posteriori variance factor and variance-covariance matrices of 
the derived coordinates, are not available. A commonly used 
procedure to evaluate the data accuracy 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 footprints. 
Such inability is a major drawback since the horizontal quality 
of the LiDAR footprints is known to be inferior to the quality of 
these points in the vertical direction. 
In the past few years, several methods have been developed for 
evaluating and/or improving LiDAR data quality by checking 
the compatibility of LiDAR footprints in overlapping strips 
(Kilian et al., 1996; Crombaghs et al., 2000; Maas, 2000; Bretar 
et al., 2004; Vosselman, 2004; Pfeifer et al., 2005). In 
Crombaghs et al. (2000), a method for reducing vertical 
discrepancies between overlapping strips is proposed. Since the 
horizontal quality of the derived point cloud is considerably 
lower than the vertical one, this approach is not sufficient to 
evaluate the overall quality of the LiDAR data. In Kilian et al. 
(1996), an adjustment procedure similar to the photogrammetric 
strip adjustment was introduced for detecting discrepancies and 
improving the compatibility between overlapping strips. The 
drawback of this approach is relying on distinct points to relate 
* Corresponding author.