de la 
je la 
o les 
caphie 
d'une 
S et 
j""Sur 
| dans 
| type 
S; En 
piéce 
fier 
e et 
Nous 
nt de 
5. L2 
Cette 
29g S0 
zone 
nt le 
^ pour 
rement 
te le 
ses 
je des 
és en 
le ces 
é ont 
LASER RANGE SCANNER SUPPORTING 3-D RANGE AND 2-D GREY LEVEL IMAGES 
FOR TUNNEL SURFACE INSPECTION 
C. Frôhlich, M. Mettenleiter and G. Schmidt 
Laboratory of Automatic Control Engineering 
Technical University of Munich, D-80290 Munich, Germany 
KEY WORDS: range sensor, range finders, cw lasers, range data, integration, tunnel inspection 
ABSTRACT 
For the survey and inspection of buildings, tunnels or underground canals, a non-tactile, robust and precise imaging of height and 
depth profiles is a basis sensor technology. For visual inspection, surface classification, and documentation purposes, however, 
additional information concerning reflectance of measured objects is necessary. High-speed acquisition of both geometric and 
visual information is achieved by means of an eyesafe laser range scanner, supporting consistent range and grey level profiles 
which are similar to video images. 
This paper reports the optical, electronic and mechanical system design of the scanner, its information processing system inclu- 
ding dedicated signal processors, and a transputer network for sensor data processing. The paper discusses experimental results of 
the laser range scanner with respect to noise as well as its long-term stability, and accuracy. Finally, results from tunnel surface 
inspection with the laser range scanner are presented. 
1. INTRODUCTION 
Range sensing is a crucial component of automation in indus- 
trial processes. To realize robotic systems [1,2] for manu- 
facturing, handling, transportation and inspection tasks, how- 
ever, it is the only way to provide the system with three di- 
mensional information [3] of its environment. The classical 
computer vision approach to range sensing is to use passive 
techniques such as stereo-vision or motion stereo. However, 
those techniques are not yet sufficiently reliable or fast 
enough to be used in many applications, most notably real- 
time robotic or inspection systems. Active sensors, which 
generate the illumination instead of using only the ambient il- 
lumination, have received increasing attention as a viable 
alternative to passive sensors, featuring direct access to range 
information in real-time. Their importance was recognized 
relatively early. For example, Besl [4] examines a wide 
variety of active range imaging technologies, comparing them 
quantitatively by evaluating a figure of merit based on range 
accuracy, field of view, and image acquisition time. A review 
of using range sensing devices in autonomous navigation of 
mobile robots can be found in Hebert ([5], Everett [6], Elfes 
[7] and Rozmann[8]. Nitzan et al. [9] describe a system inter- 
preting indoor scenes by use of range and intensity 
information from a laser ranging system. 
Through the framework of SFB 331 "Information Processing 
in Autonomous Mobile Robots", a 3-D laser range camera 
[10,11,12] was developed for indoor vehicle and robotic ap- 
plications. The laser range camera is an optical-wavelength 
radar system, and is comparable to devices built by Erim [13], 
Odetics [14], and Perceptron [15], measuring the range 
between sensor and target surface as well as the reflectance of 
the target surface which corresponds to the magnitude of the 
back scattered laser energy. In contrast to these range sensing 
devices, the laser camera under consideration is designed for 
eyesafe operation, emitting a minimum of near-infrared laser 
energy. 
For surface inspection of railway and highway tunnels how- 
ever, the performance of the range measuring system of the 
laser camera was improved with a look towards robustness, 
range accuracy, and range precision. 
This paper reports design and practical details of the laser 
range scanner for tunnel surface inspection. Chapter 2 out- 
lines the performance requirements with tunnel surface 
inspection and introduces the measurement principle of the 
sensor. Hardware design of the laser range scanner, including 
the main modules, such as the laser head with transmitter and 
receiver electronics, high frequency unit, laser beam deflec- 
tion system, and digital signal processing unit are discussed in 
Chapter 3. The signal processing unit consists of dedicated 
signal processors and a transputer system for real-time sensor 
data processing. Chapter 4 focuses on statistical evaluation of 
range data, including noise, drift over time, precision, and 
accuracy with range measurements. It discusses the influences 
of ambient light, surface material of the target, and ambient 
temperature for range accuracy and range precision. 
Furthermore, experimental results from the inspection of 
tunnel surfaces are presented. Chapter 5 concludes the paper 
by summarizing its results and gives a short outlook to future 
work in the tunnel surface inspection project. 
2. DESIGN ASPECTS 
2.1 Performance requirements 
With railway tunnel inspection, the precise measurement of 
the tunnel tube geometry, tracks and contact wires (Fig. 3a) is 
essential in order to hold tolerances for safe passing of trains. 
Furthermore, detection of possible cracks (even capillary 
cracks) in the tunnel surface is important in order to initiate 
repair work of the tunnel tube in time. For tunnel inspection, 
measurement of 360° profiles while navigating through the 
tunnel tube is more sufficient than measuring 3-D images. To 
meet these demands, high spatial resolution, accurate and 
471