www.geokno.

com A GMR Group Company

LiDAR Technology
and its Applications
…
Dr. Bharat Lohani
Professor, IIT Kanpur (on leave)
&
Exec. Director, Geokno India Pvt. Ltd.

blohani@iitk.ac.in 9450346658
blohani@geokno.com
Outline
• Specific needs of 3D data for modern/traditional
applications
• Principle of LiDAR technology
• Advantages of LiDAR data
• Comparison with other competing technologies
• Latest LiDAR data projects by Geokno
• Latest R&D on LiDAR data at IIT Kanpur

2
Applications of Topographic Data
• Cadastral Survey/Re-Survey
– Not only horizontal data but Vertical data also
• Disaster Management
– Flood Modelling and Forecast
– Cyclone Impact Analysis and Mitigation
– Landslide Hazard Zonation
• Irrigation Projects
– Command Area Survey
– Lift Irrigation Planning /Design
– Minor Irrigation Planning / Design

All these can be answered if we have accurate, precise, dense, reliable Topographic data.
3
Applications of Topographic Data…Contd
• City Surveys
– Property Solution
– Rooftop Solar Potential Mapping
• Forest Survey Requirement
– Individual Tree Mapping
– Carbon Stock Estimation
– Water Resource Planning
– Soil Erosion Susceptibility Mapping
• Road and Railway DPR generation
All these can be answered if we have accurate, precise, dense, reliable Topographic data.
4
Some of the modern applications
• Autonomous vehicles see through LiDAR
• UAV traffic management systems
• Smart city – 3D City

5
3D data requirement

Fast, Accurate, Comprehensive, Reliable, Cost-

effective method of 3D data capture required !

6
Currently there are multiple surveys with lot of redundancy and no
synchronization of data. This leads to higher cost.

MULTIPLE SURVEYS VS ONE SURVEY

CITY AS AN EXAMPLE

7
Traditional methods are not suitable
• Traditional methods
– Total Station
– GPS/GNSS
– Satellite Imageries
– Drone Photogrammetry –limited area only
• Not suitable due to
– Not applicable in inaccessible terrains – hilly terrain
– Not possible in remote areas – a surveyor cannot go there
– Too much time taking – results in delays thus cost escalation
– Errors in survey – need to revisit and re-survey and change design as
survey is wrong
– Incomplete data- data not available in forested area and shadow areas
Survey by Drone is only suitable for small and non-vegetated areas where accuracy is not important. 8
 Comparison of Technologies
Parameter for comparison TS/ GNSS Drone Satellite LiDAR
Photogrammetry
Accuracy
Speed
Completeness
Reliability of data
Cost effectiveness
Suitability for inaccessible terrain
Suitability for wooded terrain
Automation
Human dependence
Need to revisit field
Disturbance to ongoing operations
Exposure of surveyors
Suitability for small area
Suitability for large area 9
Impact of High Accuracy Data-Flood Example

10
New Technology of Aerial LiDAR

AERIAL LIDAR WITH PHOTOGRAPHS

PROVIDE COMPLETE SOLUTION

11
LiDAR technology
• Initial laser vector is transformed through a series of reference systems to
yield object coordinates in a chosen Coordinate System.

t
T/R R
z d
y d = c*t/2
GPS
c = Velocity of light
z t = Time of travel of laser pulse
x
y
IMU
z X AG = RLG X AL + X LG
x y
Laser G G
scanner X   r11 r12 r13 
x Laser X AG =  Y  RLG = r21 r22 r23 
   
pulse  Z  A  r31 r32 r33  L

12
Airborne LiDAR

GNSS Satellite
GNSS Unit

Area covered IMU

LiDAR Scanner
GNSS FOV
Ground
Station DA
H 2Φ

700m
Over 500,000 pulses
fired in a second

Ds

Laser measures range from aircraft to the point hit in ground. This range is converted to the
coordinates of the points in ground using on-board GPS and IMU. The final result is large number of
points with their known coordinates along with coloured images of entire ground. 13
LiDAR in multiple return mode
• Instrument is timed to pickup signals at certain intervals.
• First pulse to survey the top of objects while the last pulse is
used to survey the ground below.
• Intermediate pulses convey information about vertical structure
of object.

First Pulse

Intermediate
Pulses

Last Pulse

14
Initial data processing

On-board GNSS
Accurate position IMU data of
of aircraft 1 Hz aircraft 400 Hz

Reference GNSS

Aircraft trajectory Laser scanner data

and orientation range and angle

3D point cloud (E, 3 D point cloud (E,

N, H, I) as .las file Datum transfer
N, h, I)

15
Format for LiDAR data
• ASPRS format LAS
• Versions 1.0, 1.1, 1.2, 1.3, 1.4
• http://asprs.org/Committee-General/LASer-LAS-File-Format-
Exchange-Activities.html

• LASUtility to download from

• http://home.iitk.ac.in/~blohani/download.htm
Header information
LAS File point data records

18
LiDAR data example-Elevation

19
LiDAR data example- Intensity

20
Classified tree in LiDAR data.

21
LiDAR is rated as best technology for Topographic Data

ADVANTAGES OF LIDAR

22
LiDAR Vs Traditional Survey

Unlike traditional survey LiDAR captures everything—nothing is missing. 23

LiDAR technology is much better than traditional
technology

LiDAR
▪ Very fast: Over 300 sq km of data captured
in a day
▪ Exhaustive: Everything captured
▪ Fully Automated
▪ Highly Accurate with verifiable data
▪ Data can be used for multiple applications

Traditional Survey
▪ Slow: Only 1 or 2 sq km captured in a day
▪ Very little detail is captured
▪ Totally person dependent
▪ Inaccurate and no verifiable data
▪ Data cannot be used for any other
purpose
24
LiDAR Maps Under Forest/Crop Unlike
Drone, Total Station/GNSS, Satellite

Laser travels from holes in trees and measures points under trees also. Not possible in other methods.
25
LiDAR can See Under Tree/Forest

LiDAR measures
points under canopy 26
Forest Sample LiDAR Data and Ortho Image – Areas under
thick forest can be mapped easily with LiDAR
Bare Earth Model
Forest Ortho Image

All streams
mapped under
Geokno’s Telangana Project
dense forest

Railway lines, Highways, Oil and Gas Pipelines, Transmission lines that pass through heavy
vegetation can be mapped. Not possible for other techniques. 27
Researchers have found more than
60,000 hidden Maya ruins in
Guatemala in a major
archaeological breakthrough.
Laser technology was used to survey
digitally beneath the forest canopy,
revealing houses, palaces, elevated
highways, and defensive fortifications.

28
Satellite image

29
Aerial image

30
LiDAR point cloud

31
Channels extracted under forest cover

32
LiDAR Maps Un-mapable

Longitudinal section Cross section

LiDAR also maps objects as thin and inaccessible as transmission lines. Important in monitoring
transmission lines especially through thick jungles. 33
Steps in field to collect LiDAR data

PROJECT EXECUTION

34
Methodology
Preliminary Site Investigation & Area Study MoD DGCA Permit Submissions

Flight Planning
Ground Survey:
1. SoI BM identification
2. Horizontal Control (DGPS
Aerial Base Station
triangulated network)
Establishment
3. Vertical Control (Double
Tertiary Levelling)
4. Attribute collection
Ground Target Establishment

Pre-Processing Points

Check Points Data Capture

Data Pre-Processing

QA/QC using Check Point Survey

Data Post-Processing & Output generation 35

DGCA/MoD Permissions – Necessary for all Aerial Survey
projects including drone survey

DGCA

GS-III
Section, MoD

Military Air Naval Mil Ops

MHA/IB
Intelligence Intelligence Intelligence (GSGS)

36
Activities – Ground Survey
Preliminary Site Investigation & Area Study MoD DGCA Permit Submissions

Flight Planning
Ground Survey:
1. SoI BM identification
2. Horizontal Control (DGPS
Aerial Base Station
triangulated network)
Establishment
3. Vertical Control (Double
Tertiary Levelling)
4. Attribute collection
Ground Target Establishment

Pre-Processing Control Points

Check Points Data Capture

Data Pre-Processing

QA/QC using Check Point Survey

Data Post-Processing & Output generation 37

SoI Bench Mark need to be identified

Sample BM Description GTS BM located by team in Bharuch GTS BMV397 (112)

38
Ground Control Network- Master Control

▪Control network needs to be

established starting from SoI GCPs

▪Important to check the stability and

suitability of SoI BMs

▪Levelling network to be established

connecting SoI BMs to project
controls

39
Aerial LiDAR Survey Base Station identification

▪GNSS base stations need to

be established

▪Such that the aircraft is never

beyond 30 km distance from
base station

40
Activities – Flight Planning
Preliminary Site Investigation & Area Study MoD DGCA Permit Submissions

Ground Survey: Flight Planning

1. SoI BM identification
2. Horizontal Control (DGPS
triangulated network) Aerial Base Station
3. Vertical Control (Double Establishment
Tertiary Levelling)
4. Hydrographic Survey
5. Attribute collection Ground Target Establishment

Pre-Processing Points

Check Points Data Capture

Data Pre-Processing

QA/QC using Check Point Survey

Data Post-Processing & Output generation 41

Flight Path planning – Sample Flight Plan

Sample LiDAR Swath Coverage

Coverage

Flight Lines

Sample Image
Image Footprint
Coverage

42
Aerial Survey System

43
Flight Plans depends on Aerial LiDAR Platform – Riegl LMS
Q780 LiDAR with 100 MP Phase One Industrial Camera

LiDAR Sensor
Riegl LMS Q780

Camera 100 MP
Phase One Industrial

Positional System
IGI AeroControl
44
Activities – Output Generation
Preliminary Site Investigation & Area Study MoD DGCA Permit Submissions

Ground Survey: Flight Planning

1. SoI BM identification
2. Horizontal Control (DGPS
triangulated network) Aerial Base Station
3. Vertical Control (Double Establishment
Tertiary Levelling)
4. Hydrographic Survey
5. Attribute collection Ground Target Establishment

Pre-Processing Points

Check Points Data Capture

Data Pre-Processing

QA/QC using Check Point Survey

Data Post-Processing & Output generation 45

Data Pre-Processing

46
 Input Data

Data Post-Processing Input Data

Validation
Report Submission
to Client

DATA/Task
Segregation

LiDAR Data Optical Data

Ortho Process

Quicklook
Automatic Aerial Digitization of
Ortho (50cm)
Filtering / Triangulation all Planimetric
Generation for
Classification (Zone-wise) features
Whole Project

Manual Automatic / Radiometric

Manual Tie Point Topology
Inspection and Correction for
Generation Cleaning
Classification whole Project

Breakline and Quicklook

Ground Control
Hydrology Ortho Shape File
Point
feature Submission for Convertion
Measurement
Extraction approval

Classified
LiDAR Data and Bundle Final Ortho
Line Data Adjustment Generation
submission

Applying
AT Results Radiometric
Submission correction which
is approved

Digital Data Base

Bare Earth
Mosaicking and Model Generation
Model
Tiling ( DTDB & DCDB)
Extraction

Contour Final Ortho

Generation Submission

47
Example of LiDAR Data-Accuracy < 10 cm

LiDAR captures everything present on the ground-buildings, roads, transmission lines, trees,
field boundaries, utilities, bridges, culverts, existing rails, overhead wires, other assests etc.
48
Example of Simultaneously Captured
Aerial Image 10 cm GSD

Even tires on
rooftop are
seen
49
LiDAR Point Cloud - Classified

Different objects identified on LiDAR data using algorithms

50
Digital Elevation Model

Though LiDAR captures everything—the above ground objects can be removed from data
thus giving only the ground data, i.e., bare earth DEM 51
Contour Map up to 30 cm CI

Contours up to 30 cm can be easily generated from LiDAR points on ground 52

Ortho-Photograph up to 10 cm GSD

LiDAR DEM and simultaneously captured image provide accurate orthophoto and true orthophoto.
Orthophoto is like a map and one can measure directly on this. 53
GIS Layers/Topographic Map up to 1:1000

Example from Bullet Train Corridor project

From classified LiDAR data and orthophotos, the GIS layers are extracted—buildings, poles,
trees, bridges, land plot boundaries, water channels, transmission lines, utilities etc. 54
Data example from Bullet Train project

55
A large number of LiDAR flights have already been completed in India.
4,000,000 ha area captured by Geokno. A few listed here.

SUCCESSFUL PROJECTS BY GEOKNO

IN INDIA

56
Case Study: Bullet Train, Railways – Geokno saved over 6 months
for the prestigious Ahmedabad-Mumbai High Speed Rail Corridor

Topographical Map with Aerial Imagery

3D Point Cloud Data

57
Project Statistics
Project Statistics Aerial Survey Statistics
Project Length Surveyed 875 kms Aircraft type Eurocopter AS 350 B3
Hydrological Survey 283 rivers/streams Flying altitude 1300 ft / 400 m
Master GCP Points 81 Nos Aircraft Speed 80 knots
Secondary GCP Points 233 Nos Flying duration 03/Mar/17 to 18/Mar/17
LiDAR Target Points 103 Nos Flying length 1000 km
Levelling survey 1200 Kms (DT) Flying Hours 40 hrs
No of Features mapped 4,31,908 Nos
Resultant Point density 15 PPSM (RFP specified is
Elevation accuracy achieved 4 cm (9 cm) (RFP 10 PPSM)
specified accuracy is 10
cm) Raw data collected 4 TB

Positional accuracy achieved 9 cm (11 cm) (RFP No of Images captured 11500 Nos
specified accuracy is 15 No of Base-stations 13 Nos
cm)
Airport/heliport locations Ahmedabad
Topographical map Scale delivered 1:1000 (RFP specified Vadodhara
scale is 1:2500)
Surat
Orthophoto GSD 10 cm Valsad
DEM,DSM and Contour Interval 50 cm Mumbai
delivered 58
Case Study: Himalayan Greenfield Roads, Mizoram: Geokno completed
Aerial LiDAR Survey for thickly forested & remote highway corridor

Bare Earth Model of a steep valley

Road Alternative

Bridge Alternative

Cross-Section with
height of tree cover

59
Case Study: Disaster/Floods – Geokno completed flood disaster
management project in Himalayas, Uttrakhand

First Airborne LiDAR

Project for Major
Disaster Assessment in
India executed by
Geokno for
Area of Interest Survey of India
Specification Value
LiDAR Data ▪ 8 points per sq m except on slopes

Orthophotos ▪ 20 cm

Accuracy ▪ Vertical accuracy better than 20 cm

▪ Terrain: The project area is characterized by its
remote location, very high altitude, extreme Contours ▪ 50 cm Contour Interval
terrain variations, frequent high winds and DEM ▪ 50 cm grid DEM (Bare Earth)
variable ground cover (snow, rock, vegetation) Topographic ▪ 1:10,000 scale with 50 cm contour interval
▪ Summary: The project specifications call for a Map in digital form
exceptionally high data standards, across an
DTDB & DCDB ▪ As per feature based Data Model of SoI
exceptionally difficult site 60
Case Study: Irrigation, Telangana – Geokno helped Govt. of
Telangana ink pact on Godavari water projects with Maharashtra

Geokno team presenting project outcomes to

Hon’ble Chief Minister along with WAPCOS

61
Case Study: River-Interlinking, Andhra Pradesh – Geokno
completed River Interlinking project using Aerial LiDAR

Aerial LiDAR survey for linking River Godavari &

Project Title
River Penna

▪ Awarded 1,800 sq km
▪ Complete project to be referenced to Survey
Project Details of India benchmarks
▪ LiDAR scanner: Riegl LMSQ780
▪ Camera: Phase One 100 MP

Geokno CEO presenting benefits of LiDAR

Data ▪ LiDAR: 10 points per sq m
technology to Hon’ble Chief Minister
specifications ▪ Images: 10 cm GSD Orthophotos

▪ 3rd Indian state after Telangana & Rajasthan

Significant to use LiDAR technology for River Basin
project projects
achievements ▪ To use completely Indian team and Geokno
owned equipment

▪ Success of our projects in Telangana &

Project Rajasthan shows the strength of LiDAR
Learning technology and Geokno’s delivery for Aerial
LiDAR projects in India
62
Case Study: River-Interlinking, Rajasthan – Geokno completed
the Eastern Rajasthan Canal Project using Aerial LiDAR

Aerial LiDAR survey for Eastern Rajasthan Canal Project to link

Project Title
rivers Chambal, Parbati & Kalisindh
▪ Initially awarded 850 sq km
▪ Phase I of the project got extended to 2,200 sq km
▪ Further project extended by another 700 sq km for a total of
Project
2,900 sq km
Details
▪ Complete project referenced to Survey of India benchmarks
▪ LiDAR scanner: Riegl LMSQ780
▪ Camera: Phase One 100 MP
Data ▪ LiDAR: 10 points per sq m
specifications ▪ Images: 10 cm GSD Orthophotos
▪ Used completely Indian team and Geokno owned equipment
Significant ▪ Phase 1 of project data captured within one month of MoD
project clearances
achievements ▪ First major river interlinking project completed using LiDAR
technology
Current ▪ Security vetting scheduled for Phase I
project status ▪ Phase II data capture to start soon
Project ▪ Completely Indian resourced solutions can achieve faster
Learning project completions 63
Aerial LiDAR Survey for Yamuna Water Transfer
from Tejawala, Haryana to Jhunjhunu, Rajasthan

Aerial LiDAR Survey for Water Link

Project Title
between Tejawala and Jhunjhunu

▪ Project Area – 1300 sq km

Project ▪ Haryana and Rajasthan
Details ▪ LiDAR scanner: Riegl LMSQ780
▪ Camera: Phase One 100 MP

Data
▪ LiDAR: 10 points per sq m
specification
▪ Images: 10 cm GSD Orthophotos
s

▪ PDCORE, Rajasthan
Client ▪ Water Resource Department,
Rajasthan

▪ Permission obtained – in 1 month

▪ Data collection completed – in 1 week
Status ▪ Data processing – Started
▪ Delivery – 15 days from now
▪ Complete project duration – 2 months
64
Aerial LiDAR Survey for Kalisindh and Parwati Micro Lift
Irrigation Scheme - Indira Sagar Project

Aerial LiDAR Survey for Kalisindh and

Project Title Parwati Micro Lift Irrigation Scheme -
Indira Sagar Project

▪ Study Area – 9040 sq km

Project ▪ Madhya Pradesh,
Details ▪ LiDAR scanner: Riegl LMSQ780
▪ Camera: Phase One 100 MP

Data
▪ LiDAR: 10 points per sq m
specification
▪ Images: 10 cm GSD Orthophotos
s

▪ L&T Construction
Client ▪ Narmada Valley Development
Authority

▪ MOD Inspection Completed, Aerial

Status
Flying Started 65
 Case Study: Solar Rooftop Mapping – Geokno completed
India’s first project for Solar Rooftop Mapping potential project

▪ LiDAR is the used world-wide for mapping city

management and infrastructure projects including solar
rooftop potential
▪ Geokno has been awarded the India’s first project for
Solar rooftop potential mapping for Bengaluru

Source: http://www.sciencealert.com/rooftop-solar-could-provide-almost-40-percent-of-us-electricity ;
https://www.deccanchronicle.com/nation/current-affairs/070318/ 66
bengalurus-first-3-d-map-to-harness-solar-energy.html
Ortho Photo

67
Shadow Map – 01/01/2018 – 08:00

68
Shadow Map – 01/01/2018 – 09:00

69
Shadow Map – 01/01/2018 – 10:00

70
Shadow Map – 01/01/2018 – 11:00

71
Shadow Map – 01/01/2018 – 12:00

72
Shadow Map – 01/01/2018 – 13:00

73
Shadow Map – 01/01/2018 – 14:00

74
Shadow Map – 01/01/2018 – 15:00

75
Shadow Map – 01/01/2018 – 16:00

76
Shadow Map – 01/01/2018 – 17:00

77
Annual Global Horizontal Irradiance

78
Rooftops on Map Server

79
Selected Building Map Server
Sr. Attribute Value
No.

1 Perimeter 94.572
(m)
2 Area (m2) 324.964

3 Building Top 929.773

Elevation (m)

4 Building 916.984
Bottom
Elevation (m)

5 Building 12.789
Height (m)

80
Selected Building on Google Maps
Sr. Attribute Value
No.

1 Perimeter 94.572
(m)
2 Area (m2) 324.964

3 Building Top 929.773

Elevation (m)

4 Building 916.984
Bottom
Elevation (m)

5 Building 12.789
Height (m)

Mean solar insolation 1448 kWhm-2 81

Selected Building on Ortho+Google Roads
Sr. Attribute Value
No.

1 Perimeter (m) 94.572

2 Area (m2) 324.964

3 Building Top 929.773

Elevation (m)

4 Building 916.984
Bottom
Elevation (m)

5 Building 12.789
Height (m)

82
Case Study

3D CHANDIGARH

83
Case Study: Chandigarh City Mapping – Geokno is undertaking India’s
first project for SMART City Mapping using Mobile & Aerial LiDAR

84
85
86
Complete Sector Ortho with Contour

87
Complete Sector Digital Elevation Model with contour

Slope

88
Complete Sector Ortho with Contour

Zoomed Ortho with Topo Map

89
Zoomed Ortho with Topo Map

90
 Ortho - Zoom

Street View
3D Point Cloud

91
92
93
Zoomed Ortho with Street View

94
Zoomed Ortho with Street View & LiDAR Data

LiDAR Data
Building
Footprint
Floors Height

95
96
97
98
Individual Tree Mapping

FOREST MAPPING FOR CARBON

STOCK ESTIMATION

99
Raw Images- high resolution up to 5cm GSD

100
Perspective View-Colour by Intensity

101
Perspective View-Colour by Height

102
Image and LiDAR Point Cloud

103
Image and LiDAR Point Cloud

104
Processing to derive forest related parameters leading to Carbon Stock
estimation

PROCESSING STEPS

105
Digital Surface Model (DSM)- 0.5m Grid

106
Digital Elevation Model (DEM)- 0.5m Grid

107
Detailed 3D Model of the DEM of the area

108
Orthophoto- 5 cm GSD

109
Terrain Slope Map

110
Terrain Aspect Map

111
Terrain Slope Direction Map

112
Colour by Segmented Trees

113
Tree Segmented LiDAR Points-
Top View and Cross-section

114
Tree Segmented LiDAR Points-
Top View and Cross-section

115
Tree Segmented and Open Area

116
Tree Density Map

Site to identify
trees competing
for Sun Light—
possible sites for
felling.

117
Point Cloud of Individual Tree

118
Individual Tree with Details

Tree Height

Canopy
Projection
Area(CPA)

Canopy Base
Height(CBH)

Tree Location &

Ground
Elevation
119
Field Sample Sites and Correlation Matrix
• Field plot for sampling -100 m by 100 m or
specified

• Determination of field parameters using

Terrestrial Laser Scanning

Field observed parameter

– Tree height
– CBH
– Girth
– DBH
– Tree Specie

• Developing correlation between field LiDAR derived parameter

observed and LiDAR derived parameters
120
GIS Layers of Forest Area with Tree Parameters
Canopy Spread Area
Open land
Hydro-Features
Tree points

121
 Extracted Tree Parameters- GIS Database

FID Shape ID Lat Long Gnd_Elev CPA_LDR TH_LDR CBH_LDR

0 Polygon 0 18° 7' 40.681" N 81° 6' 43.237" E 146.2 10.8 15.0 7.6
1 Polygon 1 18° 7' 36.836" N 81° 6' 40.497" E 145.4 10.0 4.5 2.4
2 Polygon 2 18° 7' 37.040" N 81° 6' 40.144" E 144.7 48.2 16.5 7.3
3 Polygon 3 18° 7' 39.450" N 81° 6' 38.452" E 145.6 56.4 17.2 2.3
4 Polygon 4 18° 7' 39.175" N 81° 6' 38.326" E 145.1 1.2 5.1 1.3
5 Polygon 5 18° 7' 39.181" N 81° 6' 38.479" E 145.2 1.9 4.7 1.9
6 Polygon 6 18° 7' 35.510" N 81° 6' 40.345" E 145.9 24.5 15.2 6.4
7 Polygon 7 18° 7' 35.309" N 81° 6' 40.585" E 145.5 60.4 18.6 8.0
8 Polygon 8 18° 7' 38.587" N 81° 6' 42.885" E 145.7 3.6 7.2 4.2
9 Polygon 9 18° 7' 38.414" N 81° 6' 43.283" E 146.2 35.4 18.0 2.2
10 Polygon 10 18° 7' 38.145" N 81° 6' 42.871" E 145.9 5.4 4.0 1.9
122
Carbon Stock Estimation – Allometric Equation

• Carbon Stock = f(TH, CBH, CPA, Tree Specie)

• Allometric equations are to be developed

– Using the parameters derived from LiDAR
– With the help of existing allometric equations with FSI

123
LiDAR and Photograph together prove best solution

LAND RECORD MAPPING

SURVEY AND RESURVEY

124
Identification of field boundaries

Invisible boundaries identified

due to elevation differences

125
Identification of field boundaries
3D LiDAR View

Trees

Ortho Photo View

Fields with
boundaries
clearly visible
Mapping boundaries under foliage
• Field boundary invisible in photograph is seen in LiDAR data

127
Mapping boundaries under foliage
• Objects underneath foliage can be mapped by removing over-
ground foliage

Bare earth
generation

128
LiDAR and Dam Reservoir Analysis

129
Channel Mapping Using LiDAR Data

130
Geological Mapping Using LiDAR Data

131
 Research being pursued at IIT Kanpur
Sample papers in last few months

Kumar, B., Pandey, G., 2018 A multi-faceted CNN architecture for ISPRS, 147 (2019) 80-90
Lohani, B., automatic classification of mobile LiDAR data
and an algorithm to reproduce point cloud
samples for enhanced training
Kumar, B., Lohani, B., 2018 Development of deep learning architecture IJRS,
Pandey, G., for automatic classification of outdoor DOI: 10.1080/01431161.2018.1547929
mobile LiDAR data
Lohani, B., Singh, S., 2018 A New Approach for Determination of Solar Remote Sensing Letters, 2018 VOL. 9, NO. 7,
Chaudhary, D. and Potential Using Terrestrial Images 636–645
Nagarajan, B. https://doi.org/10.1080/2150704X.2018.145
2061
Kumar, B., Yadav, M., 2018 A two-stage algorithm for ground filtering of International Journal of Remote Sensing.
Lohani, B., Singh, A. K., airborne laser scanning data DOI: 10.1080/01431161.2018.1466074
Yadav, M., Singh, A. K., 2018 Computation of road geometry parameters Remote Sensing Applications: Society and
Lohani, B., using mobile LiDAR system Environment,
https://doi.org/10.1016/j.rsase.2018.02.003
Lohani, B. and Ghosh, S. 2017 Airborne LiDAR Technology: A Review of Proc. Natl. Acad. Sci., India, Sect. A Phys. Sci.
Data Collection and Processing Systems https://doi.org/10.1007/s40010-017-0435-9
Yadav, M., Singh A., and 2017 Extraction of road surface from mobile International Journal of Remote Sensing
Lohani, B., LiDAR data of complex road environment 38(16), 4645-4672
A.Kealy, S. Goel; . 2017 Cooperative localization of unmanned aerial ASCE Journal of Surveying Engineering, Vol.
• DST Project on Generation and Labelling of LiDAR Point Cloud for Automatic Classification through CNN approach
V.Gikas,G. Retscher, c. vehicles using GNSS, MEMS inertial and UWB 143, Issue 4 (November 2017) 132
Some results of Deep Learning Classification
realized
Oakland data set

Classes: Ground, House, Pole, Tree

Accuracy 96.5%, Kappa 91.9%

133
Some results of Deep Learning Classification
realized
Lille data set

Classes: Ground, House, Pole, Tree

Accuracy 96.2%, Kappa 92.5%

134
What about the cost?
• Cost is similar to Total
Station / GPS Survey LiDAR

• Satellite data can be

cheaper-but with poorer

Benefit
Drone

accuracy and not complete

coverage of terrain
TS
Satellite GNSS

• LiDAR most cost-effective

in terms of cost/benefit
Cost
135
 Geokno on YouTube:
https://www.youtube.com/channel/UCTcHSwRhGvusB3NxACelY3g/videos

136