Generating LiDAR data in laboratory: g y LiDAR Simulator Bharat Lohani, R K Mishra Department of Civil Engineering Department of Civil Engineering Indian Institute of Technology Kanpur Kanpur INDIA
M Motivation… i i
Are LiDAR data available ? � LiDAR data NOT available in majority of � LiDAR data NOT available in majority of countries � Lack of awareness � Lack of awareness � Security issues � Cost Bharat Lohani, IIT Kanpur India
Are LiDAR data available ? � LiDAR data NOT available for teaching � LiDAR data NOT available for teaching purposes � Readily available data � Readily available data � Data with as-desired specifications � Data with ground truth g Bharat Lohani, IIT Kanpur India
Are LiDAR data available ? � LiDAR data NOT available for research � LiDAR data NOT available for research � Data with a wide range of desired specifications � Data with complete and 100% accurate ground � Data with complete and 100% accurate ground truth Bharat Lohani, IIT Kanpur India
Solution lies in LiDAR simulator… � User creates a terrain � User creates a terrain � User chooses the flight parameters � LiDAR data are generated for created LiDAR data are generated for created terrain as if the actual LiDAR sensor had flown the terrain flown the terrain Bharat Lohani, IIT Kanpur India
Design consideration for Design consideration for simulator
Should be . . . � User friendly � User friendly � Wider distribution � Help or tutorial Bharat Lohani, IIT Kanpur India
Can simulate . . . � Generic sensor � Specific sensors � Specific sensors � ALTM � ALS � And others… Bharat Lohani, IIT Kanpur India
Should simulate trajectory as in a normal flight � 6 degrees of freedom 6 degrees of freedom Bharat Lohani, IIT Kanpur India
Should simulate earthlike surfaces Source: Optech Inc. Bharat Lohani, IIT Kanpur India
Also… � Possibility of error introduction � Possibility of error introduction � Output data available in common formats Bharat Lohani, IIT Kanpur India
Development of simulator Development of simulator
System components Integration Sensor component T Terrain i Out Input component put Trajectory Trajectory component Bharat Lohani, IIT Kanpur India
T j Trajectory component t p t Flight direction Location Attitude
Location � Location: coordinates of laser head at each firing of pulse � Location depends on Instantaneous accelerations accelerations � Instantaneous accelerations should be simulated as in a normal flight: pseudo- random simulation Bharat Lohani, IIT Kanpur India
Acceleration simulation π π ⎛ ⎞ ⎛ ⎞ J K 2 2 ∑ ∑ ∑ ∑ = + + i ⎜ ⎜ ⎟ ⎟ ⎜ ⎜ ⎟ ⎟ a A sin B ( ( id )) C cos D ( ( id )) m id ( ) X X j j j j t t k k k k t t t t ⎝ ⎝ ⎠ ⎠ ⎝ ⎝ ⎠ ⎠ T T T T = = j 1 k 1 F = Firing frequency J,K,A,B,C,D and m governing parameters parameters Bharat Lohani, IIT Kanpur India
Location simulation + = 1 + + 2 i 1 i i i 2 X X u d a d x t x t 2 X u x = Velocity in direction flight i.e. X axis V l it i di ti fli ht i X i Bharat Lohani, IIT Kanpur India
Attitude ( Roll, Pitch, Yaw ) simulation π π ⎛ ⎞ ⎛ ⎞ J K 2 2 ∑ ∑ ∑ ∑ = + + i ⎜ ⎜ ⎟ ⎟ ⎜ ⎜ ⎟ ⎟ R A sin B ( ( id )) C cos D ( ( id )) m id ( ) j j j j t t k k k k t t t t ⎝ ⎝ ⎠ ⎠ ⎝ ⎝ ⎠ ⎠ T T T T = = j 1 k 1 Bharat Lohani, IIT Kanpur India
Sensor components Sensor components Sinusoidal scan pattern p Zig-zag scan pattern
Sinusoidal scan pattern p � Let time taken to complete 1/4 th of a scan is T. � P is the numbers of points in 1/4 th of a scan. p � The maximum scan angle is ө max. ( ( ) ) π θ = θ sin t i i max 2 T 1 T where where, i t t = i i i P 0.5 Z (m) 0 -0.5 -1 160 140 140 160 160 140 120 120 100 100 80 80 60 Bharat Lohani, IIT Kanpur India Y (m) X (m)
Zig-zag scan pattern g g p max i θ θ θ = i P 1 0.5 Z (m) 0 -0.5 -1 160 140 140 140 120 120 100 100 80 80 60 Bharat Lohani, IIT Kanpur India Y (m) X (m)
Terrain component Terrain component Modeling surfaces: earthlike � Vector approach: mathematical surfaces � Raster approach with over ground objects � Fractal terrain
Vector approach: Mathematical surfaces pp Simple Surface Simple Surface • AX+BY+CZ+D=0 Example of a simple surface: 2X+5Y+10Z-100=0 (displayed in surfer) Bharat Lohani, IIT Kanpur India
Vector approach: Mathematical surfaces pp Complex Surface p • Z=A[sin(X/B)+sin(XY/BC)]+D Example of a complex surface: z=10[sin(X/25)+sin(XY/(25*50)]-300 (displayed in surfer) Bharat Lohani, IIT Kanpur India
Raster surfaces Bharat Lohani, IIT Kanpur India
Fractal surfaces Bharat Lohani, IIT Kanpur India
Integration of components Integration of components
Integration of components − − − i i i i i i X X X X Y Y Y Y Z Z Z Z = = i i i a b c Bharat Lohani, IIT Kanpur India
Error introduction in simulated data = = + + N μ N μ σ σ i i 2 X X X X ( ( , ) ) T t X X Bharat Lohani, IIT Kanpur India
Concept implementation Concept implementation
Bharat Lohani, IIT Kanpur India
Bharat Lohani, IIT Kanpur India
Bharat Lohani, IIT Kanpur India
Bharat Lohani, IIT Kanpur India
Optimal flight lines given by system p g g y y Bharat Lohani, IIT Kanpur India
User defined flight lines g Bharat Lohani, IIT Kanpur India
Bharat Lohani, IIT Kanpur India
Bharat Lohani, IIT Kanpur India
Bharat Lohani, IIT Kanpur India
Bharat Lohani, IIT Kanpur India
Simulated data and results Simulated data and results
Altitude=210m 3D raster terrain (displayed in Overlap=4% Velocity=60m/s surfer) f ) Sensor-ALS-50 Firing frequency=20KHz Scan frequency=48Hz Scan angle=40 ° Flight area=430m 430m Bharat Lohani, IIT Kanpur India
Bharat Lohani, IIT Kanpur India LiDAR data plot in plan A-A B-B
Profile A-A without error Profile A-A with error Bharat Lohani, IIT Kanpur India
Profile B-B w. r. t. flight lines g Bharat Lohani, IIT Kanpur India
LiDAR data without error Bharat Lohani, IIT Kanpur India
LiDAR data with error Bharat Lohani, IIT Kanpur India
Data with no attitude variation Bharat Lohani, IIT Kanpur India
Data with attitude variation Bharat Lohani, IIT Kanpur India
Fractal data displayed in surfer Bharat Lohani, IIT Kanpur India
Bharat Lohani, IIT Kanpur India
Bharat Lohani, IIT Kanpur India
Terrain with objects Bharat Lohani, IIT Kanpur India
Altitude=490m LiDAR data of terrain with objects Overlap=2% Velocity=60m/s Sensor-ALS-50 Firing frequency=20KHz Scan frequency=48Hz Scan angle=50 ° Flight area=640m × 460m Bharat Lohani, IIT Kanpur India
Altitude=400m Overlap=2% Effect of data density Velocity=60m/s Sensor-ALS-50 Firing frequency=20KHz Scan frequency=48Hz Scan angle=50 ° Bharat Lohani, IIT Kanpur India
Altitude=400m Overlap=2% Velocity=60m/s Sensor-ALS-50 Firing frequency=5 KHz Scan frequency=48Hz Scan angle=50 ° Bharat Lohani, IIT Kanpur India
1 Profile view of buildings 2 2 Profile view-1 Profile view-2 Bharat Lohani, IIT Kanpur India
Altitude=200m Effect of different altitude Overlap=2% Velocity=60m/s Sensor-ALS-50 Sensor-ALS-50 Firing frequency=5 KHz Scan frequency=48Hz Scan angle=50 ° Bharat Lohani, IIT Kanpur India
Altitude=100m Overlap=2% Velocity=60m/s Sensor-ALS-50 Sensor-ALS-50 Firing frequency=5 KHz Scan frequency=48Hz Scan angle=50 ° Bharat Lohani, IIT Kanpur India
Altitude=200m Effect of different scan angle Overlap=2% Velocity=60m/s Sensor-ALS-50 Sensor-ALS-50 Firing frequency=5 KHz Scan frequency=48Hz Scan angle=50 ° Bharat Lohani, IIT Kanpur India
Altitude=200m Altitude=200m Overlap=2% Velocity=60m/s Sensor-ALS-50 Firing frequency=5 Firing frequency 5 KHz Scan frequency=48Hz Scan angle=32 ° Scan angle=32 Bharat Lohani, IIT Kanpur India
Effect of different flight direction Bharat Lohani, IIT Kanpur India
Bharat Lohani, IIT Kanpur India
Applications of simulator Applications of simulator
Education � To understand: � Process of data generation � Effect of change in various parameters on data � Effect of errors on data Bharat Lohani, IIT Kanpur India
Laboratory exercises � Data with varied specifications � Data with varied specifications � Full and accurate ground truth known Full and accurate ground truth known Bharat Lohani, IIT Kanpur India
Research projects � Evaluation of Information extraction algorithms g � Assessing effect of error on performance of g algorithms � Finding optimal data specifications for an application Bharat Lohani, IIT Kanpur India
Recommend
More recommend
