Download full-text PDF

Improving strategies for the Self-Calibration of Terrestrial Laser Scanners

Presentation (PDF Available)  · October 2016with51 Reads
DOI: 10.13140/RG.2.2.26225.68960
Geodetic Week 2016, Hamburg, Germany, DOI:10.13140/RG.2.2.26225.68960
Improving strategies for the Self-Calibration of
Terrestrial Laser Scanners
Geodetic Week 2016, Hamburg
Tomislav Medic, Christoph Holst, Heiner Kuhlmann
13. October 2016
Tomislav Medic 16. October 2016 Improving Self-Calibration of the Laser Scanners Slide 1
Tomislav Medic 16. October 2016 Improving Self-Calibration of the Laser Scanners
Current Standpoint
Slide 2
Industry demands
high accuracy
Manufacturers respond with factory calibration
Problem of the end user:
Which instrument misconstructions
are calibrated? Functions?
•How many?
With which accuracy ?
How to control and recalibrate ?
www.psomas.com
www.odellengineering.com
Leica Geosystems, Walsh (2015)
Leica Geosystems (2013)
Tomislav Medic 16. October 2016 Improving Self-Calibration of the Laser Scanners Slide 3
Objectives
www.whistleralley.com
www.leica-geosystems.com
C. Holst (2015)
1. Defining optimal functional model of
mechanical misconstructions
2. Development of standardized
calibration approach for TLS
Make contribution to:
Definite answer is still
missing
Tomislav Medic 16. October 2016 Improving Self-Calibration of the Laser Scanners Slide 4
1. Optimal functional model
C. Holst (2014) C. Holst (2014) C. Holst (2014)
www.whistleralley.com
www.leica-geosystems.com C. Holst (2016)
Mechanical misconstructions:
Tomislav Medic 16. October 2016 Improving Self-Calibration of the Laser Scanners Slide 5
2. Standardized calibration approach
C. Holst (2015)
Abbas et al. (2015)
Reshetyuk (2010)
Lichti et al. (2011)
Successfully used by
many authors so far:
D. Lichti (2011)
Y. Reshetyuk (2010)
M. A. Abbas (2015)
D. García-San-Miguel
(2013)
J.L. Lerma (2014)
•…
Target-Based
Self-Calibration
Tomislav Medic 16. October 2016 Improving Self-Calibration of the Laser Scanners Slide 6
Experiment
Facts:
300 paper black & white targets
Hall dimensions: 75 x 33 x 9 m
Measurements in 2 faces
3 scanner stations
Tomislav Medic 16. October 2016 Improving Self-Calibration of the Laser Scanners Slide 7
Data Processing
Least squares based adjustment algorithm
3 similar approaches: GHM,GMM &“Unified approach”
•TLS measurements
(XYZ polar coordinates)
Target coordinates
(approximate)
Scanner position &
orientation (approximate)
Stochastic information
Measurement
corrections
Calibration
parameters
Target
coordinates
Scanner position
and orientation
Input: Output:
 ⇒ .
⇒.
Tomislav Medic 16. October 2016 Improving Self-Calibration of the Laser Scanners Slide 8
Initial Results – Functional models
2 separated sets of calibration parameters !
Empirical
parametrisation
Mechanical parametrisation Adopted from NIST
Derived from
residual plots
= 7
= 16
Before Before
After After
∆ = sin +

∆ℎ = 
tan()+
sin()+2
()+
tan()+
−
 sin +
 cos +
 cos 2ℎ + sin(2ℎ)
∆ = ()
+
cos()
+
+
 cos − sin
−
 sin +
 cos 2+
 sin(2)
∆ = 
∆ℎ = sin +
cot +
cos +
sin(3ℎ)
∆ = +
cos()
Tomislav Medic 16. October 2016 Improving Self-Calibration of the Laser Scanners Slide 9
Initial Results-Functional models
Empirical parametrisation (7)
Mechanical parametrisation (16)
measuredrange,
measuredhorizontalangle, -measuredverticalangle
∆-rangecorrection, ∆ℎ -horizontalanglecorrection, ∆ verticalanglecorrection
Tomislav Medic 16. October 2016 Improving Self-Calibration of the Laser Scanners Slide 10
Initial Results - Improvement
Distribution of measurement residuals:
Without
calibration
Empirical
param.
Mechanical
param.
Overall improvement in weighted RMSE 30% in both cases
Vertical angle
Horizontal angle
Tomislav Medic 16. October 2016 Improving Self-Calibration of the Laser Scanners Slide 11
Current Problems
Empirical Parameters
Par. Value Max.
correlation
Possible mechanical
explanation
a00.92 mm 0.16 Rangefinder offset???
b15.77 “ 0.15 Mirror tilt???
b2-5.81 “ -0.08 Beam tilt + transit tilt???
b3-1.68 “ 0.13
Horizontal encoder
eccentricity???
b4-1.18 “ 0.15
Inclination of TLS due to
movement???
c0-6.90 “ -0.23 Vertical index offset ?
c1-13.14 “ -0.23 Beam tilt, mirror tilt…. ???
Mechanical Parameters
Par. Value Max.
correlation Mechanical explanation
x1n -0.14 mm -0.59 1st component of beam offset
x1z -0.08 mm 0.92 2nd component of beam offset
x20.04 mm 0.22 Tra n s i t o ffset
x3-0.19 mm -0.62 Mirror offset
x4-101.71 “ 0.99 Vertical index offset
x63.90 “ -0.62 Mirror tilt
x73.46 “ 0.92 Transit tilt + Beam tilt (2nd)
x8x
-1.69 “ -0.41 Horizontal angle encoder eccentricity
(1st)
x8y
-4.38 “ -0.81 Horizontal angle encoder eccentricity
(2nd)
x9n -16.59 “ -0.62 Vertical angle encoder eccentricity (1st)
+ Beam tilt (1st)
x9z -151.55 “ 0.99 Vertical angle encoder eccentricity
(2nd) +Beam tilt (2nd)
x10 0.79mm -0.57 Rangefinder offset
x11a -0.87 “ 0.68 Horizontal encoder scale error (1st)
x11b 1.11 “ -0.38 Horizontal encoder scale error (2nd)
x12a 57.47 “ -0.99 Vertical encoder scale error (1st)
x12b -7.90 “ -0.76 Vertical encoder scale error (2nd)
Optimal
Parametrisation
Tomislav Medic 16. October 2016 Improving Self-Calibration of the Laser Scanners Slide 12
Current Attempts of Solving Problems
1. Defining optimal functional model of mechanical deviations
Clear mechanical interpretation, but uncorrelated?
Reusable in different measurements setups
Usable for a variety of instruments
Constant in time
Currently testing:
Different algorithms
Different datum definitions
Compensator measurements
Referent measurements of higher nominal
accuracy
Different instruments
Wunderlich et al. (2013)
www.megapixl.com
Tomislav Medic 16. October 2016 Improving Self-Calibration of the Laser Scanners Slide 13
Future Steps
Planned:
Integrate incidence angle and
intensity information in
functional/stochastic model
Integrate correlation information of
laser scanner measurements (T. Jurek,
C. Holst, H. Kuhlmann)
Integrate quality information regarding
target center estimation (J. Janßen,
C. Holst, H. Kuhlmann)
Soudarissanane et al. (2011)
Jurek (2016)
Tomislav Medic 16. October 2016 Improving Self-Calibration of the Laser Scanners Slide 14
Final Goals
Optimal functional model defined
Optimal calibration approach
Neitzel et al. (2016)
Optimal geometry for detection of mechanical misconstructions
Optimal number of scanner stations
Optimal number of targets
Optimal target type
Reference values
www.berntsen.com www.leica-geosystems.com
Tomislav Medic 16. October 2016 Improving Self-Calibration of the Laser Scanners Slide 15
Final Goals
Optimal functional model
Optimal calibration approach
Fixing permanent calibration polygon in calibration hall
Developing automatic, “user-friendly” approach
Possibly easy to reproduce
Providing all necessary information
Tomislav Medic 16. October 2016 Improving Self-Calibration of the Laser Scanners Slide 16
Thank you very much for your attention !
M.Sc. Tomislav Medic
Institut für Geodäsie und Geoinformation, Universität Bonn
Tel.: 0228/73-2623
Email: medic@igg.uni-bonn.de
Literature:
Holst, C., Kuhlmann, H. (2014): Aiming at self-calibration of terrestrial laser scanners using only
one single object and one single scan, Journal of Applied Geodesy 2014, 8(4): 295-310
Muralikrishnan, B., et al. (2015): Volumetric performance evaluation of a laser scanner based on
geometric error model, Precision Engineering 40, p. 139-150
Chow, J.C.K., et al. (2013): Improvements to and Comparison of Static Terrestrial LiDAR Self-
Calibration Methods, Sensors (2013) 13, p. 7224-7249
Reshetyuk Y. (2010): A unified approach to self-calibration of terrestrial laser scanners, ISPRS
Journal of Photogrammetry and Remote Sensing 65, p. 445-456
Lichti, D.D., Chow, J.C.K., Lahamy H.(2011): Parameter de-correlation and model-identification in
hybrid-style terrestrial laser scanner self-calibration, ISPRS Journal of Photogrammetry and
Remote Sensing 66, p. 317-326
Tomislav Medic 16. October 2016 Improving Self-Calibration of the Laser Scanner Folie 17
Initial Results
2 separated sets of calibration parameters
Empirical parametrisation NIST list
Improvement Empirical NIST
% MSE 29.69 29.01
% Range 3.30 4.40
% H. angle 19.12 19.12
% V. angle 27.92 25.38
Error plots
Muralikrishnan et al. (2014)
Tomislav Medic 16. October 2016 Improving Self-Calibration of the Laser Scanner Folie 18
Initial Results – Functional models
2 separated sets of calibration parameters !
Empirical
parametrisation
Mechanical parametrisation Adopted from NIST
Derived from
residual plots
= 7
= 16
Tomislav Medic 16. October 2016 Improving Self-Calibration of the Laser Scanners Slide 19
Current Attempts of Solving Problems
1. Defining optimal functional model of mechanical deviations
Constant in time
Reusable in different measurements setups
Usable for a variety of instruments
Uncorrelated, with clear mechanical interpretation
Currently testing:
Different algorithms
Different datum definitions
Compensator measurements
Referent measurements of higher nominal
accuracy
Different instruments
Wunderlich et al. (2013)
www.megapixl.com
This research doesn't cite any other publications.

Recommendations


Project
Development of the optimal standardized approach for the user calibration of the terrestrial laser scanners. Main focus of the study is on the panoramic type laser scanners.
Project
A proposed DFG Cluster of Excellence that focuses basic and applied research activities towards a major shift in sustainable crop production by two specific aims: multi-scale monitoring of plants a…" [more]
Article
May 2017 · Sensors · Impact Factor: 2.25
    Terrestrial laser scanner measurements suffer from systematic errors due to internal misalignments. The magnitude of the resulting errors in the point cloud in many cases exceeds the magnitude of random errors. Hence, the task of calibrating a laser scanner is important for applications with high accuracy demands. This paper primarily addresses the case of panoramic terrestrial laser scanners.... [Show full abstract]
    Article
    November 2014 · Journal of Applied Geodesy
      When using terrestrial laser scanners for high quality analyses, calibrating the laser scanner is crucial due to unavoidable misconstruction of the instrument leading to systematic errors. Consequently, the development of calibration fields for laser scanner self-calibration is widespread in the literature. However, these calibration fields altogether suffer from the fact that the calibration... [Show full abstract]
      Article
      January 2016 · Journal of Applied Geodesy
        Due to improved laser scanning technology, laser scanner based deformation analyses are presently widespread. These deformation analyses are no longer based on individual points representing the deformation of an object at selected positions. Instead, they are based on a large number of scan points sampling the whole object. This fact either leads to challenges regarding metrological aspects... [Show full abstract]
        Article
        January 2017 · Journal of Applied Geodesy
          In terms of high precision requested deformation analyses, evaluating laser scan data requires the exact knowledge of the functional and stochastic model. If this is not given, a parameter estimation leads to insufficient results. Simulating a laser scanning scene provides the knowledge of the exact functional model of the surface. Thus, it is possible to investigate the impact of neglecting... [Show full abstract]
          Article
          January 2016
            Terrestrial laser scanning was used for deformation monitoring of structures and geoscientific objects already at an early stage of its employment for engineering geodetic tasks. This paper gives an overview on current state of the art methods and fields of applications concerning the modelling of point clouds obtained with this measuring technology, by following two aims. On the one hand, the... [Show full abstract]
            Discover more