AUTOMATIC MATCHING OF TERRESTRIAL SCAN DATA USING ORIENTATION 
HISTOGRAMS 
A.G. Chibunichev, A.B.Velizhev 
Moscow State University of Geodesy and Cartography, Moscow, Russia -(agchib)@mail.ru, -(velizhev)@gmail.com 
Commission V, WG V/3 
KEY WORDS: LIDAR, Laser scanning, Point Cloud, Registration, Orientation, Transformation 
ABSTRACT: 
Terrestrial laser scanners allow to obtain accurate 3d model of the object as a point cloud. Real objects should be scanned with 
different instrument positions and then scan data is registered into the single coordinate system. In practice the process of 
registration is performed using manual or semi-automated registration techniques. First fully automatic methods were introduced 
about fifteen years ago. Many of the methods based on well-known iterative closest point algorithm (ICP) permit to obtain the exact 
solution of a problem. Nevertheless, the problem is difficult to solve due to necessity of information regarding the coarse relative 
orientation parameters of point clouds. In this paper authors present the algorithm based on usage of orientation histogram to solve a 
problem without any information concerning scan positions. Experimental studies of the algorithm have demonstrated its efficiency 
at any values of point clouds relative orientation. 
1. INTRODUCTION 
Terrestrial laser scanners have been widely adopted in problem 
solution deals with documentation of historic buildings and 
monuments, industrial complexes monitoring, industrial and 
city units mapping. The 
result of scanning is performed by high density 3d model of the 
point cloud that precisely describes the surface of object survey. 
The scan is the point cloud obtained at immobile scanner 
position. The points of each scan have been obtained in the 
coordinate system of scanner arbitrary oriented in the space. To 
process several scans simultaneously we need to know relative 
orientation parameters for all scans obtained. 
In practice before scanning process starts artificial markers are 
placed on the surface of surveyed object. Results of scanning 
these markers’ are used to integrate the scans into a common 
coordinate system. Some commercially available programs 
allow realizing this operation in semi-automatic or automatic 
modes. Nevertheless, theoretically there is enough information 
coded in scans itself to carry out its integration. Therefore the 
majority of software solutions have an option for manual 
selection of corresponding scan areas that can be used to solve a 
problem. The problem of finding the registration parameters 
without artificial markers has wide range of features in 
automation area. That is why many research groups are 
involved in the problem. Quite lot solutions for point cloud 
registration with different range of accuracy have been 
introduced during last 10 years. 
2. RELATED WORK 
Iterative closest point algorithm (ICP) described in the article 
(Besl et al., 1992) becomes a standard solution for the problem 
of automatic scans joining. The main advantage of ICP 
algorithm is its simple implementation but the main 
disadvantage is necessity to know accurate values of first 
approximation of the scans orientation parameters. 
But generally there is no any information concerning 
approximate relative point cloud orientation. Therefore the ICP 
algorithm is applied on the final stage of solution to improve 
precision. Scans pre-alignment is the major problem in its 
automatic registration process. Multiple solutions of this 
problem are proposed and analyzed by many research groups. 
(Feldmar et al., 1996) have described a pre-alignment algorithm 
based on the corresponding point matching. (Ripperda et al., 
2005) use a normal distribution transform for cloud alignment. 
(Liu et al., 2005) have introduced a solution based on the 
automatic scan segmentation. Authors applied scan 
segmentation based on difference between the normal vectors. 
Matching of the segments was performed using a special 
matching tree. However, the same segments of different scans 
often has a large difference in shape, therefore its matching is 
very difficult. 
(Dold et al., 2004; Vanden Wyngaerd et al., 1999) have 
described a coarse pre- alignment by matching the same planes 
on the scans. 
To detect angular orientation of scans the author of the paper 
(Dold, 2005) used an extended Gaussian images (EGI) 
matching algorithm on different levels of detail. (Makadia et al., 
2006) used EGI to make orientation histograms that were 
compared with correlation function. In this paper the pre 
alignment algorithm has been introduced to improve the method 
using simplification of mathematics. 
3. ORIENTATION HISTOGRAM 
601 
EGI that were introduced in (Horn, 1984) article are one of the 
way for presentation of objects 3D shape. EGI creation was 
made by superposition of normal vectors of the object surface