IMPROVEMENT OF THE GEOSPATIAL ACCURACY OF MOBILE TERRESTRIAL LiDAR DATA ...

295 Pages·2016·9.61 MB·English

IMPROVEMENT OF THE GEOSPATIAL ACCURACY OF MOBILE
TERRESTRIAL LiDAR DATA

MICHAEL LESLAR

A DISSERTATION SUBMITTED TO
THE FACULTY OF GRADUATE STUDIES
IN PARTIAL FULFILLMENT OF THE REQUIREMENTS
FOR THE DEGREE OF
DOCTOR OF PHILOSOPHY

GRADUATE PROGRAM IN EARTH AND SPACE SCIENCE
YORK UNIVERSITY
TORONTO, ONTARIO

NOVEMBER 2016

© MICHAEL LESLAR, 2016
ABSTRACT

Many applications, such as topographic surveying for transportation engineering, have
specific high accuracy requirements which MTL may be able to achieve under specific
circumstances. Since high rate, immersive (360° FOV), MTL is a relatively new device
for the collection and extraction of survey data; the understanding and correction of
errors within such systems is under researched. Therefore, the goal of the work presented
here is to quantify the geospatial accuracy of MTL data and improve the quality of MTL
data products.

Quantification of the geospatial accuracy of MTL systems was accomplished through the
use of residual analysis, error propagation and conditional variance analysis. Real data
from two MTL systems was analyzed using these methods and it was found that the
actual errors exceeded the manufacturer’s estimates of system accuracy by over 10mm.
Conditional variance analysis on these systems has shown that the contribution by the
interactions among the measured parameters to the variances of the points in MTL point
clouds is insignificant. The sizes of the variances for the measurements used to produce a
point are the primary sources of error in the output point cloud.

Improvement of the geospatial accuracy of MTL data products was accomplished by
developing methods for the simultaneous multi-sensor calibration of the system’s
boresight angles and lever arm offsets, zero error calibration, temperature correction, and
both spatial and temporal outlier detection. Evaluation of the effectiveness of these
techniques was accomplished through the use of two test cases, employing real MTL
ii
data. Test case 1 showed that the residuals between a control field and the MTL point
cloud were reduced by 4.4cm for points located on both horizontal and vertical target
surfaces. Similarly, test case 2 showed a reduction in the residuals between control
points and MTL data of 2~3cm on horizontal surfaces and 1~2cm on vertical surfaces.
The most accurate point cloud produced through the use of these calibration and filtering
techniques occurred in test case 1 (27mm ± 26mm). This result is still not accurate
enough for certain high accuracy applications such as topographic surveying for
transportation engineering (20mm ± 10mm).
iii
ACKNOWLEDGMENTS
I would like to acknowledge the financial assistance provided by the Natural Sciences
and Engineering Research Council of Canada (NSERC) and the Ministry of Transport
Ontario (MTO). I would also like to acknowledge Optech Incorporated for providing the
equipment and LiDAR Data used in this dissertation. Finally, I would like to
acknowledge the significant amount of assistance and advice I received from both Dr.
Jian-Guo Wang, P.Eng and Dr. Baoxin Hu, P.Eng.
iv
TABLE OF CONTENTS
ABSTRACT ........................................................................................................................ ii
ACKNOWLEDGMENTS ................................................................................................. iv
TABLE OF CONTENTS .................................................................................................... v
LIST OF TABLES ............................................................................................................. ix
LIST OF FIGURES .......................................................................................................... xii
LIST OF ABREVIATIONS ............................................................................................ xix
1. Introduction ................................................................................................................. 1
1.1 Motivation ................................................................................................................. 1
1.2 Objectives ................................................................................................................. 3
1.3 Outline....................................................................................................................... 4
2. The Basics of Mobile Time-Of-Flight LiDAR and Literature Review ...................... 7
2.1 Components of Mobile Time-Of-Flight LiDAR Systems ........................................ 8
2.1.1 The Laser ........................................................................................................... 8
2.1.2 The Range Finder ............................................................................................. 12
2.1.3 The Deflection Unit ......................................................................................... 14
2.1.4 The Control and Data Recorder ....................................................................... 16
2.1.5 The DG System ................................................................................................ 17
2.1.6 The Correction Station ..................................................................................... 22
2.2 Errors in Mobile Time-Of-Flight LiDAR Systems ................................................. 23
2.2.1 Instrument Errors ............................................................................................. 26
2.2.1.1 Random Errors .......................................................................................... 26
2.2.1.2 Timing Errors ............................................................................................ 28
2.2.1.2.1 Calibration/Malfunction ..................................................................... 30
2.2.1.2.2 Time Walk ......................................................................................... 34
2.2.1.2.3 Returned Intensity .............................................................................. 37
2.2.1.4 Zero and Scale Error ................................................................................. 39
2.2.1.5 Temperature Drift ..................................................................................... 41
2.2.1.6 Mixed Pixels ............................................................................................. 42
2.2.1.7 Dynamic Track Error ................................................................................ 43
2.2.1.8 Velocity Errors .......................................................................................... 44
2.2.1.9 Scanner Assembly Balance ....................................................................... 45
2.2.2 Environmental Errors ....................................................................................... 45
2.2.2.1 Atmospheric Propagation Errors ............................................................... 45
2.2.2.2 Adverse Weather Conditions .................................................................... 47
2.2.2.3 Other Atmospheric Conditions ................................................................. 48
2.2.2.4 Interfering Radiation ................................................................................. 51
2.2.2.5 Scanner Instability and Vibration ............................................................. 51
2.2.3 Instrument Position Errors ............................................................................... 52
2.2.3.1 Trajectory Errors ....................................................................................... 52
2.2.3.2 Boresight and Lever Arm Errors ............................................................... 53
2.2.3.2.1 Heading Errors ................................................................................... 54
2.2.3.2.2 Roll and Pitch Errors .......................................................................... 57
2.2.3.2.3 System Lever Arm Errors .................................................................. 60
2.2.3.2.4 Methods for Computing Boresight Angles and Lever Arms ............. 61
v
2.2.4 Target Errors .................................................................................................... 65
2.2.4.1 Object Reflectance .................................................................................... 65
2.2.4.2 Laser Beam Incidence Angle .................................................................... 67
2.2.4.3 User Pick Errors ........................................................................................ 68
2.3 Outlier Detection ..................................................................................................... 69
2.3.1 Distribution-Based Outlier Detection .............................................................. 70
2.3.2 Depth-Based Outlier Detection ........................................................................ 71
2.3.3 Distance-Based Outlier Detection .................................................................... 72
2.3.4 Clustering-Based Outlier Detection ................................................................. 72
2.3.5 Density-Based Outlier Detection ..................................................................... 73
3. Error Analysis of MTL Systems ............................................................................... 76
3.1 Introduction ............................................................................................................. 76
3.2 The Mathematical Models ...................................................................................... 77
3.2.1 LiDAR to Geocentric Coordinate Conversion ................................................. 77
3.2.2 Error Propagation Analysis .............................................................................. 81
3.2.3 Conditional Variance Analysis ........................................................................ 82
3.3 Comparing MTL to Control .................................................................................... 84
3.4 Sensitivity Testing .................................................................................................. 88
3.5 Error Propagation .................................................................................................... 90
3.6 Conditional Variance Analysis ............................................................................... 96
3.7 Discussion ............................................................................................................. 102
3.8 Summary ............................................................................................................... 105
4. Calibration of MTL Systems .................................................................................. 106
4.1 Zero Error of an MTL Sensor ............................................................................... 106
4.1.1 Introduction .................................................................................................... 106
4.1.2 Mathematical Model and System Setup ........................................................ 107
4.1.3 Experimental Results ..................................................................................... 110
4.1.4 Field Verification of Results .......................................................................... 114
4.2 Temperature Changes in MTL Sensors ................................................................ 118
4.2.1 Introduction .................................................................................................... 118
4.2.2 Establishing the Amount of Temperature Walk in an MTL Sensor .............. 121
4.2.2.1 Experiment 1 ........................................................................................... 121
4.2.2.2 Experiment 2 ........................................................................................... 123
4.2.3 Estimating Temperature Corrections ............................................................. 126
4.2.4 Evaluating the Results .................................................................................... 128
4.3 Calibration of MTL Sensors to the DG System .................................................... 131
4.3.1 Introduction .................................................................................................... 131
4.3.2 The Mathematical Models ............................................................................. 133
4.3.2.1 Finding Boresight Angles and Lever Arms from Ground Control
(BLAGC) ............................................................................................................ 133
4.3.2.2 Finding Boresight Angles and Lever Arms from Opposing Data Strips
(BLAOS) ............................................................................................................. 134
4.3.2.3 Finding Boresight Angles and Lever Arms from Data from Two Sensors,
Collected Concurrently (BLATS) ....................................................................... 136
4.3.2.4 Correcting Point Selections..................................................................... 138
4.3.2.5 Adjustment Statistics .............................................................................. 142
vi
4.3.3 Implementation and Testing .......................................................................... 145
4.3.3.1 The Calibration Site ................................................................................ 145
4.3.3.2 Utility Development ................................................................................ 148
4.3.3.3 Determining the Minimum Number of Control Points to Incorporate into
the Adjustment .................................................................................................... 150
4.3.4 Results and Discussion .................................................................................. 157
4.4 Summary ............................................................................................................... 164
5. Outlier Detection and Removal from MTL Data .................................................... 167
5.1 Introduction ........................................................................................................... 167
5.2 Mathematical Models ............................................................................................ 169
5.2.1 The α-β-γ Kalman Smoother ......................................................................... 169
5.2.2 Quadratic Polynomial Surface Fitting (PSF) ................................................. 171
5.3 Implementation ..................................................................................................... 176
5.4 Algorithm Tests .................................................................................................... 179
5.4.1 Case Study 1 – Vaughan, Ontario, Canada .................................................... 183
5.4.1.1 Test 1 – Appropriate Window Size for the Time Series Approach ........ 185
5.4.1.2 Test 2 – Appropriate Sample Size for the Spatial Series Approach ....... 186
5.4.1.3 Test 3 – Maximum Number of Outliers Detectable by the Routines ..... 188
5.4.1.4 Test 4 –Using Commercial Software to Detect Outliers ........................ 189
5.4.2 Case Study 2 – Pontarlier, France .................................................................. 190
5.4.2.1 Test 1 – Appropriate Window Size for the Time Series Approach ........ 192
5.4.2.2 Test 2 – Appropriate Sample Size for the Spatial Series Approach ....... 193
5.4.2.3 Test 3 – Maximum Number of Outliers Detectable by the Routines ..... 194
5.4.2.4 Test 4 –Using Commercial Software to Detect Outliers ........................ 195
5.4.3 Case Study 3 – Washington D.C., U.S.A. ...................................................... 196
5.4.3.1 Test 1 – Appropriate Window Size for the Time Series Approach ........ 197
5.4.3.2 Test 2 – Appropriate Sample Size for the Spatial Series Approach ....... 199
5.4.3.3 Test 3 – Maximum Number of Outliers Detectable by the Routines ..... 200
5.4.3.4 Test 4 – Using Commercial Software to Detect Outliers ....................... 201
5.5 Discussion ............................................................................................................. 202
5.5.1 Results of Window Size Determination in the Time Series Approach to Outlier
Detection ................................................................................................................. 205
5.5.2 Results of Patch Size Determination in the Spatial Series Approach to Outlier
Detection ................................................................................................................. 207
5.5.3 Results of Maximum Number of Outliers Detectable by the Routines ......... 209
5.5.4 Results of Using Commercial Software to Detect Outliers ........................... 211
5.6 Summary ............................................................................................................... 212
6. Improving the Accuracy of MTL Point Cloud Data ............................................... 214
6.1 Procedure for Improving the Accuracy of MTL Point Clouds ............................. 215
6.2 Application to Real MTL Data ............................................................................. 217
6.2.1 Data of a Commercial Office Building .......................................................... 217
6.2.1.1 Assessing the Quality of the Data before Processing ............................. 220
6.2.1.2 Zero Error and Temperature Correction ................................................. 226
6.2.1.3 Calibrating the LiDAR to the DG System .............................................. 227
6.2.1.4 Removing Outliers .................................................................................. 229
6.2.1.5 Assessing the Quality of the Data after Processing ................................ 231
vii
6.2.2 Data of a Typical Street Scene ....................................................................... 235
6.2.2.1 Assessing the Quality of the Data Before Processing ............................. 238
6.2.2.2 Zero Error and Temperature Correction ................................................. 242
6.2.2.3 Calibrating the LiDAR to the DG System .............................................. 243
6.2.2.4 Removing Outliers .................................................................................. 245
6.2.2.5 Assessing the Quality of the Data after Processing ................................ 247
6.3 Summary ............................................................................................................... 251
7. Conclusions and Recommendations for Future Work ............................................ 252
7.1 Conclusions ........................................................................................................... 252
7.2 Recommendations for Future Work...................................................................... 257
REFERENCES ............................................................................................................... 259
Appendix A ..................................................................................................................... 271
Roll (θ) .................................................................................................................. 271
1
Pitch (θ ) ................................................................................................................ 272
2
Heading (θ) ........................................................................................................... 272
3
X Lever Arm(l ) .................................................................................................... 273
X
Y Lever Arm(l ) .................................................................................................... 273
Y
Z Lever Arm(l ) ..................................................................................................... 273
Z
POS Roll (r) ........................................................................................................... 274
p
POS Pitch ( ) ........................................................................................................ 274
h
POS Heading ( ) .................................................................................................... 275
POS X (x ) ........................................................................................................... 276
pos
POS Y ( y ) .......................................................................................................... 276
pos
POS Z (z ) ........................................................................................................... 276
pos

viii
LIST OF TABLES
Table 2.1: Specifications for the Applanix LV DG system [41]. ..................................... 18
Table 2.2: Specifications for the SPAN GNSS/INS combined systems [42]. .................. 20
Table 2.3: Specifications for the iXblue INS-GNSS systems [43]. .................................. 21
Table 2.4: Zero error estimates for various time-of-flight and phase based STL[15]. ..... 40
Table 2.5: Range changes due to temperature drift from literature sources. .................... 42
Table 2.6: Coefficients for the indexes of dry air and water vapour under standard
conditions [68]. ............................................................................................... 49
Table 3.1: Expected and ideal uncertainties in MTL parameters. .................................... 93
Table 4.1: Manufacturer’s zero error estimates for Lynx Mobile Mapper sensor
SN131104. .................................................................................................... 111
Table 4.2: Zero and scale errors for Lynx sensor SN131104 at a laser PRF of 500kHz. 111
Table 4.3: Zero and scale errors for Lynx sensor SN131104 at a laser PRF of 250kHz. 111
Table 4.4: Zero and scale errors for Lynx sensor SN131104 at a laser PRF of 125kHz. 112
Table 4.5: Zero and scale errors for Lynx sensor SN131104 at a laser PRF of 75kHz. . 112
Table 4.6: Zero errors for Lynx sensor SN131104 at a laser PRF of 500kHz. ............... 113
Table 4.7: Zero errors for Lynx sensor SN131104 at a laser PRF of 250kHz. ............... 113
Table 4.8: Zero errors for Lynx sensor SN131104 at a laser PRF of 125kHz. ............... 113
Table 4.9: Zero errors for Lynx sensor SN131104 at a laser PRF of 75kHz. ................. 114
Table 4.10: Maximum and minimum data values for ranges measured during temperature
testing. Results are for experiments conducted on April 7th, 2011............. 125
Table 4.11: Maximum and minimum data values for ranges measured during temperature
testing. Results are for experiments conducted on June 6th, 2011. ............. 126
Table 4.12: Comparison of LiDAR measured ranges to total station measured ranges
before and after temperature compensation was applied. This trial was
conducted on October 22nd, 2011. ............................................................... 129
Table 4.13: Comparison of LiDAR measured ranges to total station measured ranges
before and after temperature compensation was applied. This trial was
conducted on March 18th, 2012. .................................................................. 130
Table 4.14: Generic boresight and lever arm values for a Lynx Mobile Mapper system.
..................................................................................................................... 146
Table 4.15: Manufacturer provided boresight and lever arm values for the Lynx Mobile
Mapper system used in testing. ................................................................... 146
Table 4.16: ECEF control coordinates established at the target site using Total Station.
..................................................................................................................... 147
Table 4.17: Boresight and lever arm values with their standard deviations for data from a
Lynx Mobile Mapper, as provided by the manufacturer. ............................ 157
Table 4.18: Calculated boresight and lever arm values with standard deviation error
estimates for data from a Lynx Mobile Mapper point cloud. ...................... 160
Table 5.1: Specifications for point cloud sections collected in a parking lot in Vaughan
Ontario and used in algorithm testing. .......................................................... 183
Table 5.2: Best results from trials conducted using algorithm (a) on point clouds A1, B1
and C1. .......................................................................................................... 188
Table 5.3: Best results from trials conducted using algorithm (b) on point clouds B1 and
C1. ................................................................................................................. 188
ix
Table 5.4: Best results from trials conducted using algorithm (c) on point clouds A1, B1
and C1. .......................................................................................................... 188
Table 5.5: Results from trials conducted using algorithm (a) preceding algorithm (b) on
point clouds B1 and C1. ................................................................................ 189
Table 5.6: Results from trials conducted using Polyworks IMSurvey’s reject outliers
routine on point clouds A1, B1 and C1. ........................................................ 190
Table 5.7: Specifications for point cloud sections collected in a lumber yard in Pontarlier,
France and used in algorithm testing. ........................................................... 190
Table 5.8: Best results from trials conducted using algorithm (a) on point clouds A2 and
C2. ................................................................................................................. 195
Table 5.9: Best results from trials conducted using algorithm (b) on point clouds A2 and
C2. ................................................................................................................. 195
Table 5.10: Results from trials conducted using algorithm (a) preceding algorithm (b) on
point clouds A2 and C2. .............................................................................. 195
Table 5.11: Results from trials conducted using Polyworks IMSurvey’s reject outliers
routine on point clouds A2 and C2. ............................................................. 196
Table 5.12: Specifications for point cloud sections collected on a street in Washington
D.C., U.S.A and used in algorithm testing. ................................................. 197
Table 5.13: Best results from trials conducted using algorithm (a) on point clouds A3 and
C3. ............................................................................................................... 200
Table 5.14: Best results from trials conducted using algorithm (b) on point clouds A3 and
C3. ............................................................................................................... 201
Table 5.15: Results from trials conducted using algorithm (a) preceding algorithm (b) on
point clouds A3 and C3. .............................................................................. 201
Table 5.16: Results from trials conducted using Polyworks IMSurvey’s reject outliers
routine on point clouds A3 and C3. ............................................................. 202
Table 6.1: Zero error as calculated by the manufacturer. ............................................... 226
Table 6.2: Zero error calculated from the method in Chapter 5. .................................... 227
Table 6.3: Control data used with Sensor 1. ................................................................... 228
Table 6.4: Control data used with Sensor 2. ................................................................... 228
Table 6.5: Boresight and lever arm values and one sigma standard deviations for both
sensors before correction. ............................................................................. 228
Table 6.6: Boresight and lever arm values and one sigma standard deviations for both
sensors after correction. ................................................................................ 228
Table 6.7: Points removed from each sensor's point clouds by the outlier filter based on
the α-β-γ Kalman smoother. ......................................................................... 229
Table 6.8: Points that failed to return a result from the outlier filter based on the α-β-γ
Kalman smoother and were therefore indeterminate. ................................... 230
Table 6.9: Zero error as calculated by the manufacturer. ............................................... 242
Table 6.10: Control data used with Sensor 1. ................................................................. 243
Table 6.11: Control data used with Sensor 2. ................................................................. 243
Table 6.12: Boresight and lever arm values and one sigma standard deviations for both
sensors before correction. ............................................................................ 244
Table 6.13: Boresight and lever arm values and one sigma standard deviations for both
sensors after correction. ............................................................................... 244
x