Flying spot laser triangulation scanner using

lateral synchronization for surface

profile precision measurement

Hanlin Zhang, Yongjie Ren,* Changjie Liu, and Jigui Zhu

State Key Lab of Precision Measuring Technology and Instruments, Tianjin University, Tianjin 300072, China
*Corresponding author: [email protected]

Received 15 April 2014; revised 1 June 2014; accepted 1 June 2014;

posted 2 June 2014 (Doc. ID 210179); published 3 July 2014

High-speed surface profile measurement with high precision is crucial for target inspection and quality
control. In this study, a laser scanner based on a single point laser triangulation displacement sensor and
a high-speed rotating polygon mirror is proposed. The autosynchronized scanning scheme is introduced
to alleviate the trade-off between the field of view and the range precision, which is the inherent
deficiency of the conventional triangulation. The lateral synchronized flying spot technology has excel-
lent characteristics, such as programmable and larger field of view, high immunity to ambient light or
secondary reflections, high optical signal-to-noise ratio, and minimum shadow effect. Owing to
automatic point-to-point laser power control, high accuracy and superior data quality are possible when
measuring objects featuring varying surface characteristics even in demanding applications. The
proposed laser triangulation scanner is validated using a laboratory-built prototype and practical con-
siderations for design and implementation of the system are described, including speckle noise reduction
method and real-time signal processing. A method for rapid and accurate calibration of the laser triangu-
lation scanner using lookup tables is also devised, and the system calibration accuracy is generally
smaller than
0.025 mm. Experimental results are presented and show a broad application prospect
for fast surface profile precision measurement. © 2014 Optical Society of America
OCIS codes: (120.0120) Instrumentation, measurement, and metrology; (120.5800) Scanners;
(150.1488) Calibration; (150.5670) Range finding.
http://dx.doi.org/10.1364/AO.53.004405

1. Introduction inspection and reverse engineering, in recent years.

Measurement of the surface profile plays an impor- Blais has given an excellent overview over 20 years
tant role in industrial environment for both the con- of range sensor development, describing the most
trol of manufacturing processes and for final product important ranging principles that are now commer-
acceptance. However, the most widely used and reli- cially available. The range sensor accuracy varies
able surface measuring techniques either involve from 0.01 to 0.3 mm depending on measurement
contact, such as stylus instruments, or are subject range and working principle for triangulation-based
to limited depth of measurement and small working commercial systems [2]. A single point laser triangu-
distance, such as confocal laser scanning microscopes lation sensor is well developed for surface profile
[1]. Triangulation-based optical metrology has been measurement with integration of a measuring
widely used in industrial applications, such as target machine [3].
To achieve a significant speed increase, the most
straight-forward approach is the single spot laser
1559-128X/14/204405-08$15.00/0 scanner shown in Fig. 1. The laser beam is deflected
© 2014 Optical Society of America by a mirror mounted on a mechanical scanner, such

10 July 2014 / Vol. 53, No. 20 / APPLIED OPTICS 4405

 highly reflective targets, such as machined parts,
can be falsely interpreted by the two-dimensional
image sensor as true data points. The third problem
is that the horizontal and vertical fields of view and
resolution have a fixed relationship partly deter-
mined by the sensor geometry. This makes it diffi-
cult, for example, to have a small width of field
but a very large depth of field when measuring
narrow deep grooves. As light from the whole scene
enters the camera lens, the fourth problem is that the
system is sensitive to ambient light.
An interesting scanning mechanism proposed by
Rioux [6] allows the implementation of a compact
laser scanner without reducing resolution and field
of view. The design is based on the longitudinal syn-
chronization of the projection and imaging optical
paths. One side of a double-sided coated mirror is
Fig. 1. Working principle of the single spot laser scanner.
used to deflect the projecting laser beam while the
other side collects the reflected light from the scene.
as a galvanometer or a mirror coupled to a motor This kind of sensor has an accuracy of approximately
shaft. In this case, the scanning angle is varied 0.1 mm at the working distance of 150 mm–0.7 cm at
according to a predefined field of view. By simple maximum range of 1000 mm [4]. However, it aims at
trigonometry the two-dimensional coordinates of large depth of view, typically 100–1000 mm, and the
the measuring point can be calculated. However, in measurement accuracy is unsatisfactory for short-
the conventional triangulation setup, a compromise distance applications.
among the field of view, measurement uncertainty Another technique using a lateral synchronized
and shadow effects must be considered [4]. laser scanner where the instantaneous field of view
The other classical method consists of projecting a is perpendicular to the scanning direction of the spot
laser line generated by spreading the laser beam was demonstrated by Oomen and Verbeek [7] for
with a cylindrical lens on the target, and coordinates robot arc welding. In their system, two small mirrors
of the points on the projected line can be obtained on a single oscillating shaft are used for synchroniza-
using a two-dimensional charge coupled device tion of projection and reflected laser beams. The sen-
(CCD) array [5], as shown in Fig. 2. The laser stripe sor accuracy of the distance measurement is 0.3 mm
sensors have been used widely in many domains as for a maximum field of view of 60 mm × 80 mm.
they are simple, inexpensive, relatively high preci- However, the system measurement speed and accu-
sion, and easy to control. One advantage of this type racy could be better if they took into consideration
of laser sensor is that it contains no moving parts and the dynamic balance of the scanning mechanism.
can be extremely robust. In this paper, a flying spot laser triangulation
Although easy to implement, there are potential scanner based on a single point laser triangulation
significant drawbacks with this approach. First, if displacement sensor and a high-speed rotating poly-
the laser line covers a surface area with different gon mirror has been developed to provide a real-time,
reflective characteristics it is very difficult to find high-speed, and superior data quality profile mea-
a setting of laser power and camera integration time surement and the potential for measurement on a
to get high-quality data on both surface types. A sec- wide range of materials and colors. The rest of the
ond major problem associated with this laser stripe paper is organized as follows: Section 2 describes
sensor is that second reflections on the surface of the laser triangulation scanner system model and
the calibration method. The experimental setup
and calibration results are shown in Section 3 to
verify the system feasibility. Accuracy assessment
of system calibration is detailed in Section 4 to
evaluate the system performance with a practical
application. The paper is concluded with an outlook
on future work in Section 5.

2. System Configuration

A. System Principle and Measurement Procedure

The technology behind the flying spot laser scanner
is single point laser triangulation [2]. Figure 3 illus-
trates the general geometrical principle of the simple
Fig. 2. Working principle of the laser stripe sensor. laser triangulation sensor.

4406 APPLIED OPTICS / Vol. 53, No. 20 / 10 July 2014

Fig. 3. Schematic diagram of the single point laser triangulation
displacement sensor based on Scheimpflug principle.

The focused laser beam emitted from the laser

diode is projected onto the diffuse surface to be mea-
sured. A portion of the light scattered from the laser
spot is collected by an objective lens and then imaged
onto a linear position sensor. The sensor surface is Fig. 4. Schematic diagram of the flying spot laser triangulation
tilted to maintain the Scheimpflug condition such scanner using lateral synchronization.
that for a given illumination direction, any point
along the laser beam path will be in-focus on the optomechanical complexity is minimal compared
position sensor. The displacement variation Δz along with a tenfold increase in performance.
the projected laser beam is detected by calculating Measurement for a single point is trigged by pulses
the position change Δp of the imaging spot on the from the optical rotary encoder coupled to the
linear position sensor according to the triangulation rotating polygon mirror. Position change Δp of the
principle, and the relationship between these laser spot on the linear position sensor is calculated
two variables is nonlinear and can be established for a given exposure time and laser power level.
through a calibration procedure. Knowing Δp and the pulse number from the rotary
For a two-dimensional profile measurement, a fast encoder, the two-dimensional coordinates of the
and stable scanning mechanism with good dynamic measured point can be obtained after the system
balance is required. The measurement is performed calibration.
by sweeping the triangulation plane of the single
point laser sensor shown in Fig. 3 over an angle B. System Calibration Method
using a rotating polygon mirror. The imaging The precision of the measurement system relies on
and projection optical paths are mechanically the accuracy of calibration and the ability of the sys-
synchronized, as shown in Fig. 4. The flying spot tem to retain the calibrated accuracy. Nonparametric
technology indicates that the points contained in calibration methods are well adapted to small
the output line are captured sequentially, rather reconstruction volumes and to the modeling of local
than simultaneously. distortions that can include mirror surface defects,
The flying spot laser triangulation scanner using optical and geometrical distortions, and other nonli-
lateral synchronization has many advantages over nearities that may be difficult to identify or model.
the laser stripe sensor mentioned previously. Laser
power and camera integration time can be easily
and automatically controlled to optimize the re- 1. Determining Measurement Point Coordinate
ceived signal level as only a single point measure- in Z Axis
ment is made at any time. Combined with limited For convenience, as shown in Fig. 5(a), the Z axis lies
instantaneous field of view, the system is extremely parallel to the center of the system field of view, and
immune to ambient light, and also the chance of the X axis is perpendicular to it in the system scan-
secondary reflections and false points is low. The ning plane. The laser scanning beam is indexed by
depth of field for this laser scanner is defined by the pulse number from the rotary encoder, for
the linear position sensor and its associated optics. example, the projecting laser beam li . A reference
The field of view is defined by the scanning mecha- plane perpendicular to the Z axis is placed at equal
nism. These two are completely unrelated and so intervals in the system field of view with accurate Z
can be designed independently to suit a specific coordinate. For each interval position, the position
application. change Δp related to the intersection point Pij of
These benefits combine to give a considerably laser beam li and the reference plane is noted, and
more accurate measuring system with excellent the Z coordinate of point Pij is known [8]. Linear in-
raw data quality. The extra expense of additional terpolation is done between two adjacent calibrated

10 July 2014 / Vol. 53, No. 20 / APPLIED OPTICS 4407

Fig. 5. Schematic diagram of the calibration system. (a) Determining measurement point coordinate in Z axis, and (b) calibration of the
incident laser beam equations.

points for each laser beam li to obtain the Z coordi- adaptive laser power control algorithm based on
nate of the measuring point. pulse-width and pulse-height modulation is built
as the power collected by the imaging lens varies sig-
2. Calibration of the Incident Laser Beam nificantly depending on the distance to the target,
Equations surface characteristic of the target, and the incident
angle of the laser beam.
The X coordinate of the measuring point is deter- Unlike smart CCD sensors found in commercial
mined by the equation of the incident laser beam laser triangulation sensors using closed loop control
li in the defined coordinate system O − ZX. The equa- to adjust the power of the transmitting laser accord-
tions of the scanning laser beams are calibrated ing to the amount of reflected light from the target
directly and used when measuring to eliminate the [10], this flying spot laser triangulation scanner
error introduced by the traditional analytical system employs multiple scans to perform a real measure-
model [9]. The reference plane is tilted forming ment at the compromise between reduced measure-
an intersection point O with the X axis, i.e., the ori- ment speed and higher measurement accuracy,
gin of the defined coordinate system, as shown in which is detailed in the next section.
Fig. 5(b). For intersection point Pi0 , the Z coordinate
is calculated by image position change Δp and linear
interpolation, as mentioned above, then the X 2. Real-Time Data Processing
coordinate can be obtained by the equation of the The data processing chain is shown in Fig. 6 for real-
intersection line of the reference plane and the sys- time high-speed range measurement. The pulse gen-
tem scanning plane. The equation of the intersection erated by the rotary encoder triggers the camera to
line m0 at this instant can be written as acquire a line of image for a predetermined amount
of time, and the laser illumination intensity is modu-
x  kz; k  tan90° − θ; (1) lated at the same time. The acquired image data are

where θ is the inclination angle, and its value is

known. Now the inclined reference plane moves
along the Z axis with an accurate interval
Δj j  1; 2…M. Similarly, the equation of the inter-
section line mj at this instant can be written as

x  kz − Δj ; k  tan90° − θ: (2)

Now there are M  1 known points which can be

obtained for the laser beam li as the inclined refer-
ence plane moves along the Z axis. Then the equa-
tions of the incident laser beams can be obtained
by fitting the intersection points with a line in a
least-squares sense.
C. Signal Processing for Accurate Laser Spot Position
Extraction
1. Precise Laser Power Control
To assure the system accuracy over a wide dynamic Fig. 6. Laser power control and data processing timing diagram.
range and high signal-to-noise ratio, a precise α is the scanning angle.

4408 APPLIED OPTICS / Vol. 53, No. 20 / 10 July 2014

transferred to a buffer area during the linear CCD this study. A detailed calculation procedure is
readout process for further processing before the described by Kielhofer et al. [18].
next pulse from the rotary encoder arrives. This pro-
cedure is repeated at every instant that a pulse from 3. Experimental Setup and Calibration Results
the rotary encoder is received until a scanning line is
completed. A. Design and Fabrication of the Laser Triangulation
The nonscanning area included at the beginning Scanner
and end of the scanning line Σ is left for laser spot A laser diode module with wavelength of 660 nm was
position calculation of the current scanning line or used for its very large depth of focus and an output
for camera exposure time and laser illumination power of 5 mW in a small package at the compromise
intensity estimation of the next new scanning line. of increased laser speckle noise. A pair of achromatic
doublets with equivalent focal length of 50 mm were
3. Speckle Reduction Methods and Laser Spot used to reduce laser spot image distortion caused by
Position Detection lens aberrations and with aperture diameter of
10 mm for speckle reduction. The linear CCD image
Reflected laser light from the target results in sensor with 1024 pixels and 14 μm pixel size was
speckle that alters the Gaussian profile and limits used in this design. The image sensor exposure time
the accuracy of calculating the laser spot peak posi- was set to 50 μs max to assure the sampling rate of
tion [11]. Speckle noise arises when speckle elements 10 kHz at least for distance measurement. A simple
on the surface illuminated by the laser spot are laser displacement sensor subsystem can be built
larger when compared to the wavelength of the laser based on the Scheimpflug condition with the triangu-
source. According to Baribeau and Rioux [12], the lation angle of 18°.
uncertainty associated with locating the laser spot The scanning subsystem consisted of a 12 facet
peak position can be modeled as high-speed polygon mirror with mirror size of
64 mm × 8 mm driven by a direct current motor
λf and an optical rotary encoder with 2048 pulses per
σp  p





; (3) revolution. Four hundred pulses from the rotary
D cosβ 2π
encoder after quadrature decoding when the polygon
mirror had rotated one facet were used in this design.
where β is the Scheimpflug angle of the linear CCD, λ The image processing took place offline for experi-
is the wavelength of the laser source, f is the effective mental purpose in a computer with a frame grabber.
focal length of the imaging lens, D is the lens A specific developed field programmable gate array
diameter. Such choosing small aperture numbers (FPGA) board synchronized the rotary encoder sig-
for the objective is effective in reduction of the laser nals with the frame grabber and did the laser diode
speckle effect, which improves depth discrimination power modulation.
but at the expense of depth of focus. However, the Based on the aforementioned configurations, a
Scheimpflug condition makes this influence mini- flying spot laser triangulation scanner has been de-
mum. Another substantial method is to use incoher- signed, fabricated, and assembled, as shown in Fig. 7.
ent light that reduces speckle noise associated with The designed laser scanner has 70 mm measurement
lasers and consequently produces a homogeneous range with approximately 65 mm stand-off distance,
image. However, the depth of view will be reduced and the scan width at stand-off distance is approxi-
unless an autofocusing method is used when mately 88 mm.
measuring.
An alternative is altering the distance between the B. Calibration Procedure and Results
target and the sensor, which is most convenient with As soon as the prototype has been assembled, a cal-
this flying spot laser scanner [13]. Combined with a ibration procedure was performed which is based on
proper imaging processing algorithm, the influence a precision translation mechanism. Figure 8 shows
of speckle on measurement accuracy is under the schematic of the calibration system based on
control [14]. the Abbe’s principle. The surface of the ceramic
To increase the resolution of the laser scanner, it is gauge block has a matte color with Lambertian prop-
crucial to obtain the position of the laser spot on the erties. The gauge block had been mounted on a pre-
linear image sensor to subpixel accuracy. Many cision rotation stage and aligned to be perpendicular
approaches to subpixel spot peak detection are pos- to the direction of motion that defined the Z axis. The
sible and summarized by Fisher et al. [15,16]. Aware laser scanner was set up so that the gauge block, at
of the laser spot image resulting from lens distortion, its center of travel, lay at the center of the laser
laser beam asymmetry, laser beam incident angle scanner field of view and perpendicular to the laser
to the target, surface texture [17] and laser speckle, scanner scanning plane.
and benefiting from the developed precise laser Measurements of the gauge block flat surface with
power control method, a robust algorithm to different distances from the laser scanner were car-
detect reproducible spot positions called a Fourier- ried out to evaluate the repeatability. The laser spot
descriptor-based phase detector method is used in positions of 400 laser beams were calculated 30 times

10 July 2014 / Vol. 53, No. 20 / APPLIED OPTICS 4409

Fig. 7. Design and fabrication of the laser scanner. (a) Laser triangulation scanner system setup, and (b) photo of the fabricated scanner,
(1) laser diode, (2) folding mirror, (3) polygon mirror, (4) objective lens, (5) linear CCD, (6) rotary encoder.

at each distance with optimized laser power level.

The measurement standard deviations from a dis-
tance of 70 mm, shown in Fig. 9, indicate that the
peak-to-peak value of the variance among the mea-
surement results is better than 0.1 pixel, and similar
results can be obtained over the system field of view
as a result of the proposed adaptive laser power
control method.
The calibration procedure consisted of acquiring
several profiles of the surface of the gauge block
at perfectly known positions with the Renishaw
XL-80 laser interferometer as reference. Calibration
of the Z axis was done with the gauge block
perpendicular to the motion axis of the translation Fig. 9. Measurement standard deviations of laser spot positions
mechanism. Figure 10 shows the recorded image for 400 laser beams from a working distance of 70 mm.
position as the laser scanner was scanning across
the gauge block at a range of 0–70 mm in steps of
2 mm. The fall in triangulation gain with the target
range is shown by the decreasing separation of the
image position curves.
Calibration of the equations of the incident laser
beams is done with the gauge block inclined at the
angle θ. The inclination angle was set 20° using

Fig. 8. Experimental setup to calibrate the laser scanner using Fig. 10. Recorded laser spot image position raw data in steps of
the Renishaw XL-80 laser interferometer system. approximately 2 mm.

4410 APPLIED OPTICS / Vol. 53, No. 20 / 10 July 2014

 Fig. 12. Deviation of measurement results from its median value
among the whole profile for the 30 acquisitions. The short line and
Fig. 11. Calculated equations of the laser scanning beams. Not
the dot in the box represent the median and the mean value,
all laser beams are drawn for clarity.
respectively.

the precision rotation stage, and the same calibration

procedure was repeated as mentioned above. The in-
cident laser beam equations using the least-squares
method with eight intersection points are drawn in
Fig. 11 for an intuition purpose.

4. System Performance Evaluation and Application

It is worth mentioning that determining and
expressing the accuracy of the laser scanner after cal-
ibration is difficult as the calibrated points are not
equally spaced, and the accuracy of the measuring
point is nonlinear in both axes.
A. Accuracy Evaluation of Calibration
To evaluate the accuracy of the designed calibration
method, the surface of the gauge block used in cali-
bration was measured at different displacements
from the laser scanner in steps of 2 mm. At each
position, 10 measurement results for all 400 laser
beams were averaged, noted, and compared with
the reference using a laser interferometer. Figure 12
shows a boxplot of the deviation of results from its
median value along the whole profile for the 30 ac-
quisitions. It can be seen that the system calibration
accuracy is generally smaller than
0.025 mm.
B. Profile Measurement of a Manufactured Object
A two-dimensional profile of a manufactured speci- Fig. 13. Profile measurement of a manufactured object. (a) Photo
men made of Invar alloy had been measured by the of the specimen, and (b) measurement results from a working
developed laser scanner from a working distance of distance of 70 mm.
70 mm. The measurement results shown in Fig. 13(b)
indicate that the height of the measured profile is
278 μm with the least-squares method, and the 5. Conclusion
deviation is only 26 μm when compared to the refer- A high-speed real-time flying spot laser triangula-
ence of 304 μm using a coordinate measuring tion scanner using lateral synchronization is pro-
machine. posed. The autosynchronized scanning scheme is
The fluctuation of the measurement data shown in introduced to alleviate the trade-off between the field
Fig. 13(b) is largely due to the surface properties of view and the range precision, which is the inherent
resulting from the material and manufacturing deficiency of the conventional triangulation. The
process [17]. lateral synchronized flying spot technology has

10 July 2014 / Vol. 53, No. 20 / APPLIED OPTICS 4411

excellent characteristics, such as programmable field References
of view, high immunity to ambient light or secondary 1. J. Coupland, R. Mandal, K. Palodhi, and R. Leach, “Coherence
reflections, high optical signal-to-noise ratio, and scanning interferometry: linear theory of surface measure-
auto laser speckle reduction. Owing to automatic ment,” Appl. Opt. 52, 3662–3670 (2013).
point-to-point laser power control, high accuracy 2. F. Blais, “Review of 20 years of range sensor development,”
J. Electron. Imaging 13, 231–240 (2004).
and superior data quality are possible when measur- 3. F.-J. Shiou and M.-J. Chen, “Intermittent process hybrid mea-
ing objects featuring varying surface characteristics surement system on the machining centre,” Int. J. Prod. Res.
even in demanding applications. The method for 41, 4403–4427 (2003).
rapid and accurate calibration of the laser triangula- 4. F. Blais, M. Rioux, and J.-A. Beraldin, “Practical considera-
tion scanner using lookup tables is also devised. tions for a design of a high precision 3-D laser scanner sys-
tem,” Proc. SPIE 0959, 225–246 (1988).
The designed flying spot laser triangulation scan- 5. L. Xia, P. Chen, Y. Wang, L. Zhou, and X. Luo, “Three-
ner is validated using a laboratory-built prototype in dimensional measuring technique for surface topography
calibration and experiments. The results demon- using a light-sectioning microscope,” Appl. Opt. 51, 1162–
strate that the measuring repeatability is within 1170 (2012).
6. M. Rioux, “Laser range finder based on synchronized scan-
0.1 pixel, and the measuring accuracy is better than ners,” Appl. Opt. 23, 3837–3844 (1984).
0.03 mm when measuring a manufactured specimen, 7. G. L. Oomen and W. J. Verbeek, “A real-time optical profile
which shows a broad application prospect for fast sensor for robot arc welding,” Proc. SPIE 0449, 62–73
surface profile precision measurement, particularly (1984).
in the industrial field of target inspection and quality 8. F. Blais and J. Beraldin, “Calibration of an anamorphic
laser-based 3D range sensor,” Proc. SPIE 3174, 113–122
control. (1997).
For three-dimensional profile measurement, the 9. X. Li, H. Zhao, Y. Liu, H. Jiang, and Y. Bian, “Laser scanning
third dimension can be added by allowing the trans- based three dimensional measurement of vegetation canopy
lation stage to transport the target through the sys- structure,” Opt. Lasers Eng. 54, 152–158 (2014).
tem field of view, or can be operated in combination 10. J. Beraldin, J. Domey, L. Gonzo, F. Blais, M. Rioux, A. Simoni,
M. Gottardi, F. De Nisi, D. Stoppa, and F. Comper, “Optimized
with very accurate pose reference systems like position sensors for flying-spot active triangulation systems,”
measuring arms or tracking systems. in Proceedings of IEEE Conference on 3-D Digital Imaging
Although the designed system is less accurate and Modeling (IEEE, 2003), pp. 28–36.
when compared to a typical triangulation-based com- 11. R. G. Dorsch, G. Häusler, and J. M. Herrmann, “Laser triangu-
mercial laser probe from the company of Kreon with lation: fundamental uncertainty in distance measurement,”
Appl. Opt. 33, 1306–1314 (1994).
an accuracy of 0.025 mm for a similar measurement 12. R. Baribeau and M. Rioux, “Influence of speckle on laser range
range [2], the lateral synchronized flying spot finders,” Appl. Opt. 30, 2873–2878 (1991).
technology behind the designed laser scanner will 13. W. Dremel, G. Hausler, and M. Maul, “Triangulation with
guarantee a more promising measurement result, large dynamical range,” Proc. SPIE 0665, 182–187 (1986).
especially in demanding applications. Also it should 14. G. Yang, C. Sun, P. Wang, and Y. Xu, “High-speed scanning
stroboscopic fringe-pattern projection technology for three-
be pointed out here that, as shown in Fig. 10, only dimensional shape precision measurement,” Appl. Opt. 53,
about 500 pixels of the linear CCD sensor are 174–183 (2014).
included in the range calculation. System resolution 15. R. Fisher and D. Naidu, “A comparison of algorithms for sub-
and accuracy will be improved when choosing an pixel peak detection,” in Image Technology: Advances in Image
imaging lens with a larger focal length. Processing, Multimedia and Machine Vision, D. Jorge and
L. C. Sanz, eds. (Springer, 1996), pp. 385–404.
Future work will deal with the development of an 16. F. Blais and M. Rioux, “Real-time numerical peak detector,”
integrated, compact prototype for profile measure- Int. J. Prod. Res. 11, 145–155 (1986).
ment in real industrial environments without the 17. K. Žbontar, M. Mihelj, B. Podobnik, F. Povše, and M. Munih,
aid of a PC and further experiments on specimens “Dynamic symmetrical pattern projection based laser
triangulation sensor for precise surface position measurement
of different materials and colors.
of various material types,” Appl. Opt. 52, 2750–2760
The financial support from 863 High-technology (2013).
Research and Development Program of China under 18. S. Kielhofer, T. Bahls, F. Hacker, T. Wusthoff, and M. Suppa,
“DLR VR-SCAN: A versatile and robust miniaturized laser
No. 2012AA041205 and National Natural Science scanner for short range 3D-modelling and exploration in
Funds for Distinguished Young Scholar under robotics,” in Proceedings of IEEE Conference on Intelligent
No. 51225505 is gratefully acknowledged. Robots and Systems (IEEE, 2011), pp. 1933–1939.

4412 APPLIED OPTICS / Vol. 53, No. 20 / 10 July 2014