sentation for 3D 
U-RI-TR-97-47, 
ISA. 
Techniques for 
of Optical 3-D 
a. 
' Registration of 
9-256. 
997. Estimating 
y a Frequency 
Optical 3D 
erland. 
on of absolute 
nal of Optical 
HIGH RESOLUTION DIGITAL IMAGE ORIENTATION 
USING LASER SCANNER DATA 
Leandro Bornaz (*), Andrea Lingua (*), Fulvio Rinaudo (*) 
(*) Politecnico di Torino — Dipartimento di Georisorse e Territorio 
C.so Duca degli Abruzzi, 24 — 10129 TORINO 
Tel. +39.011.564.7659 / 7659 / 7700 Fax. +39.011.564.7699 
E-mail: bornaz@polito.it; lingua@polito.it; rinaudo@polito.it; 
  
KEY WORDS: LIDAR, Photogrammetry, Close Range, Architecture, Software 
ABSTRACT: 
The investigation of archaeological and architectural goods can offer great advantages thanks to the reconstruction of 3D models for 
more effective object analysis. For this purpose it is necessary to derive a dense digital elevation model (DTM) that is useful for 3D 
modelling. DTM surveying is traditionally performed using profiling total-stations or derived from photogrammetric stereo-plotting. 
In recent years, laser scanner technology has also been developed for non industrial applications and new instruments have been 
produced that satisfy the requirements of cultural heritage recording. Thousands of points can be surveyed in a few seconds, with 
accuracy that is adequate for particular applications, to build a dense DTM that can be directly compared with a great number of 
representation techniques and which can be integrated with more traditional survey methods. 
Laser scanner instruments used in terrestrial application can usually acquire a digital RGB image of the object. The resolution of this 
image is the equivalent of the laser scanning resolution: a pixel of radiometric data (grey levels or RGB) corresponds to an acquired 
3D point, creating a low resolution digital image that is not useful for a correct photographic description of the object. This 
information needs further integration with high resolution digital images acquired using a digital camera. 
This paper analyses the application possibilities of laser scanner surveying for an accurate orientation of high resolution digital 
images. For this purpose, a set of software tools have been implemented for: 
- automatic reflecting marker recognition on the inside of the laser scanner acquisition (point set 1); 
- automatic reflecting marker recognition on the inside of the digital images using the least squares matching (point set 2); 
- automatic linking between point set 1 and point set 2; 
- estimating the image orientation parameters by means of bundle block solution. 
1. INTRODUCTION When the quality (both metric and radiometric) of the image 
produced by the laser scanner is not sufficient, digital images 
The use of laser scanner devices in architectural and are supplied trought the use of traditional digital cameras which 
archaeological applications is, today, one of the most widely 
investigated techniques. 
The enormous number of points which can be surveyed in few 
minutes offers many opportunities to record and document 
cultural heritage activities. 
The quality analysis of the acquired points and the removal of 
the measurement noises can be considered at a good level of 
investigation; the achieved results (procedures and algorithms) 
are currently in use in practical surveying applications. 
The direct use of those points to extract 3D models of the 
surveyed object is one of the research activities still underway: 
the basic studies have not yet been concluded and the 
potentiality offered by a dense DTM generated using the laser 
scanner devices, has not yet been completely investigated. 
The integration of laser scanner techniques and digital 
photogrammetry can be considered one of the most interesting 
topics of present day research activities. 
In recent years some interesting applications have been 
developed (e.g. 3D realistic model build-up, true orthophoto 
production, 3D navigation). The use of laser scanner data 
always leads to a speeding up of the traditional 
photogrammetric procedures and in many cases offers new 
opportunities and new survey products. 
It is well known that some laser scanner devices can produce an 
RGB image acquired during the point acquisition phase. 
In some cases these images are of low resolution and sometimes 
the orientation parameters are not known with the accuracy that 
is necessary for a correct join between the point data set and the 
image. 
operate outside the laser scanner. 
In all cases high resolution digital images, previously oriented 
or calibrated (in the case of non metric camera) using traditional 
orientation or calibration algorithms, are employed. 
The control point coordinate measurement is usually performed 
by means of traditional topographic surveying or manual 
extraction from the point data set produced by the laser scanner. 
     
Figure 1. An image acquired by the Riegl LMS Z210 (a) and the 
same image recorded by the Nikon D1X high resolution digital 
camera (b) 
—321—