SSRG International Journal of Civil Engineering ( SSRG – IJCE ) – Volume 3 Issue 1 January 2016

Lidar-ALTM and Its Application for Urban

Planning and Infrastructure by Developing 3D
City Model
Sheena A D*, Dr. C. Udhayakumar**
*
Research Scholar, Institute of Remote Sensing, Department of Civil Engineering,
Anna University, Chennai, India.
**
Associate Professor, Institute of Remote Sensing, Department of Civil Engineering,
Anna University, Chennai, India.

Abstract ALTM (Airborne laser Terrain Mapping) is a recent

In this developing era of modern trend of new technology for capturing data of the topography
urbanization and industrialization there is an of the earth surface features. It became feasible
emergent need for proper city planning systems. 3D through the availability of lasers with Inertial
models are digital representations of the Earth’s Measurement Unit (IMU) and the Global Positioning
surface and related objects belonging to urban areas System (GPS).
on the earth surface. 3D models are real world
representation useful in applications such as 3D A typical LiDAR system is operated from a
visualization, planning the city in Infrastructure aircraft, a helicopter or a spacecraft. The setup rapidly
development, Information system for tourism, transmits pulses of laser which travel to the surface,
Intelligent transportation systems, Environmental where they are reflected back. The return pulse is
aspects, Disaster Management, public rescue converted from photons to electrical impulses and
operations, real estate market, utility management, received by a high-speed data recorder. Using the
Military operations, Training of officers, Simulation formula for the speed of light, time intervals from
of new buildings, Updating and keeping cadastral transmission to collection are easily derived. Time
data, monitoring change detection and virtual intervals are then converted to distance based on
reality. positional information obtained from GPS receivers
LiDAR data (Light Detection and Ranging) and the on-board Inertial Measurement Unit (IMU)
is a recent new technology for obtaining the earth’s which constantly records the attitude of the aircraft.
surface objects mainly for 3D building visualization. The data can be gathered during the day or night,
ALTM (Airborne Laser Terrain Mapping) data when weather and even in small rains due to its active
combined with digital orthophotos can be used to system. LiDAR systems collect positional (x, y) and
create highly detailed Digital Surface Models (DSMs) elevation (z) data at pre-defined intervals and the
and DEM from which Digital 3D City Models are resulting LiDAR data is of very dense point clouds.
obtained. The accuracy of the LiDAR data is a function of the
Various techniques and conditions were flying height, laser beam diameter, the quality of the
used to extract the objects and accuracy level is GPS or IMU data, and post-processing methods.
examined. Research in 3D GIS helps to analyse the Accuracies of less than 1 m can be achieved using
replica of real world and the related issues. Here, in LiDAR technology.
this paper the 3D objects are developed using ArcGIS
B. Objects in 3D GIS
and also by using algorithm techniques.
3D objects are x,y,z coordinate features
Keywords - 3D model, City model, Visualization, representing on the earth surface. These are extracted
Infrastructure, Disaster Management by using algorithms and conditions.
C. Need for the study
I. INTRODUCTION
Accurate and timely 3D spatial information
A. Lidar Technology about buildings in urban areas is needed as the basis
LIDAR (Light Detection And Ranging) is an to assist decision making in understanding, managing,
optical remote sensing technology that measures and planning the continuously changing environment.
properties of scattered light to find range of a distant Construction of new neighbourhoods, transportation
target in the form of laser pulses. Its application is networks, and drainage systems are a few situations
widely in archaeology, geography, geology, in which information about buildings is essential to
geomorphology, seismology, forestry, Urbanization, the planning process. However, useful quantitative
Town planning, remote sensing, Geomatics and spatial information cannot be obtained directly from
Atmospheric physics. existing data sources and must rely on extraction

ISSN: 2348 – 8352 www.internationaljournalssrg.org Page 28

 SSRG International Journal of Civil Engineering ( SSRG – IJCE ) – Volume 3 Issue 1 January 2016

techniques. Conventional field survey methods are It is to define a proper rule to identify
time-consuming and costly. Planning maps are used meaningful image objects owing to the fact that the
widely in various urban applications because of their object identification is influenced significantly by the
high accuracy, although producing and updating them knowledge and the quality of the data. Urban trees
are time-consuming and hard to automate. tend to be confused with buildings because of similar
heights of trees and low-rise buildings.
Recently available high-resolution remotely
sensed data have provided the potential to achieve Therefore, how to extract urban buildings
effective and efficient building information through accurately and efficiently from LiDAR data is one of
automated extraction methods. Aerial photographs the main aspect in current work. With this extraction
are a popular data source for building extraction of buildings, 3D model is generated to meet the need
owing to their high resolution. However, they often of solving the risks and issues in changing
cover selected areas only and lack multiple times of environment. The general methodology comprises
coverage for the same area, making it difficult to steps required to develop 3D objects.
update the building database once created. High-
resolution commercial satellite image data have been III. PREPARATION OF DSM
used increasingly in building extraction, but the METHODOLOGY
issues of shadowing and distortion can affect the Digital Elevation Model (DEM) generated
resulting accuracy to a certain degree of extend. A using Minimum Filtering Technique was used to
new data source, Light Detection and Ranging eliminate the elevation influence from the digital
(LiDAR) data, is used in this project work, which surface model (DSM) derived from the LiDAR data
provides land-surface elevation information (height) so that a normalized height model could be
by emitting a laser pulse and providing high vertical computed.
and horizontal resolutions of less than 1 m.
POINT FILE
II. GENERAL METHODOLOGY INFORMATION

ALTM Data
Ortho images
(High resolution)
3D Point clouds LAS TO
MULTIPOINT
(X,Y,Z)

MULTIPOINT
DEM generation

Extracting objects INTERPOLATIO

N TECHNIQUE -
NEAREST
NEIGHBOURHO
Model 3D objects OD

Digital Photos of
building
DSM FOCAL
STATISTICS
Generating MINIMUM
Texture Images FILTERING
TECHNIQUE
DEM
Flood Parameters
3D city modeling
NH= DSM-DEM

Web
Application Flood Impacts
OFF-TERRAIN
HEIGHTS
3D Walkthrough

Figure 2. Flowchart for generating DSM and DEM

Figure 1. Flowchart for General Methodology

ISSN: 2348 – 8352 www.internationaljournalssrg.org Page 29

 SSRG International Journal of Civil Engineering ( SSRG – IJCE ) – Volume 3 Issue 1 January 2016

A. Point File Information time and the calculation is repeated until the entire
The LAS file format is a public file format image has been filtered and a "new" image has been
for the interchange of LIDAR data between vendors generated. By varying the calculation performed and
and customers. This binary file format is an alternative the weightings of the individual pixels in the filter
to proprietary systems or a generic ASCII file window, filters can be designed to enhance or
interchange system used by many companies. The suppress different types of features. Here, minimum
LAS file format maintains the information specific to filtering is used to find out the minimum value of
the LIDAR nature of the data while not being overly kernel window in the neighbourhood to be considered
complex. as centre pixel for DEM extraction or Bare Earth
B. LAS to Multipoint Model.
This customized tool imports one or more F. Normalized Height Model
files in LAS format, into a new multipoint feature Normalized height model is derived from
class. Supported LAS file format versions are 1.0 and Digital surface model minus Digital Elevation Model.
1.1.
The .las point files are converted to Multipoints 3D NHM = DSM – DEM
using the syntax in script. 
Syntax:
LASToMultipoint_3d <input; input>
<in_feature_dataset> <out_feature_class>
{average_point_spacing} {class_code; class_code...}
{return;return...} {keyword {name}; keyword
{name}...} {input_coordinate_system} {file_suffix}

C. Interpolation Technique
Interpolation is a method of constructing
new data points within the range of a discrete set of
Figure 3. Concept of Developing NHM
known data points. Nearest neighbour resampling
uses the digital value from the pixel in the original
image which is nearest to the new pixel location in IV. ACQUISITION OF 3D OBJECTS FROM
the corrected image. This is the simplest method and ACQUIRED SENSOR DATA
does not alter the original values. The Focal Statistics
Automatic building extraction from LIDAR
tool gives control over the neighbourhood type and
Data is obtained by modelling without physically
statistics to be calculated.
going to the exact location or object. From airborne
x = (width of the neighbourhood + 1)/2
LIDAR data, digital surface model (DSM) can be
y = (height of the neighbourhood+1)/2
generated and then the objects higher from the ground
are automatically detected from the DSM. The
D. Digital Surface Model buildings, geometric characteristics such as size,
Digital Surface Models (DSM) are widely height and shape information are then used to
used in the earth sciences. It provides information for separate the buildings from other objects. The
various geological studies and other applications. extracted building outlines are then simplified using
There are a number of methods for automatic DSM an orthogonal or edge detection algorithm to obtain
generation, each of which has its own strengths and better effect.
weaknesses, none of which are perfect. Even though
there are plenty of algorithms to date which can A. Processing Of NHM or NDSM Data
generate the DSMs, it is a computationally complex Methodology
calculation and does tend to take some time to These are the step by step approach to generate
complete. In order to achieve faster DSMs, an 3D models from NHM.
algorithm was implemented on a graphics processing
unit. 1) Building Extraction Technique
In this project, the function of feature
extraction in the ArcGIS model builder was used to
E. Minimum Filtering Technique perform image segmentation. It includes a series of
A common filtering procedure involves
procedures for building extraction:
moving a 'window' of a few pixels in dimension (e.g.
3x3, 5x5, etc.) over each pixel in the image, applying
i. Segmentation: Image segmentation is the process of
a mathematical calculation using the pixel values
partitioning images into segments by grouping
under that window, and replacing the central pixel
neighbouring pixels with similar characteristics.
with the new value. The window is moved along in
both the row and column dimensions one pixel at a

ISSN: 2348 – 8352 www.internationaljournalssrg.org Page 30

 SSRG International Journal of Civil Engineering ( SSRG – IJCE ) – Volume 3 Issue 1 January 2016

ii. Merge: This process is used for solving the over d) Form Factor: Form Factor is the shape
segmentation problem and grouping small segments measure that compares the area of the polygon with
together. the square of the total perimeter. The form-factor
value of a circle is 1, and the value of a square is π/4.
iii. Threshold technique. These conditions are given as threshold factor.

Form factor = (4 x π x area)/(total perimeter)2



B. Accuracy Asseessment
NHM The accuracy of the extraction result was
assessed using the planning map as the ground truth.
Condition: Value < 0
True Value : 0; False Value :
The resulting building image was compared with a
CONDITION NHM rasterized planning map. The evaluation
measurements used were widely accepted for
building extraction (Lee et al., 2003; Shufelt, 1999),
CORRECTED NHM which categorized all pixels into four types as a result
of comparing two images pixel by pixel.
1) True positive (TP): Both the extraction and the
reference indicate that a pixel belonged to a building.
MAX. AND MIN.
CLOSING AREA
2) True negative (TN): Both the extraction and the
reference indicated that a pixel belonged to the
background.
3) False positive (FP): The extraction incorrectly
DIVIDE/3 identified a pixel as belonging to a building.
4) False negative (FN): The extraction did not
correctly identify a pixel that truly belonged to a
INT
building.

Branching factor = FP/TP 

RASTER TO POLYGON
Miss factor = FN/TP

CALCULATE AREA<
10 Detection percentage = 100 × TP/(TP + FN) 
ELIMINATE 3 
ITERATIONS Quality percentage = 100 × TP/(TP + FP + FN) 
FORM FACTOR 
THRESHOLD
V. RESULTS FOR SAMPLE DATA OF
NIAGARA
3D MODEL ACCURACY A. NHM
ACCESSMENT

Figure 4. Flowchart for processing NHM

a) Corrected NHM: A correction process was

needed to filter out abnormal points based on the
following criteria:
Condition:

If IV of NHM < 0, then IV of NHM = 0;

Else IV of NHM = IV of NHM 

Figure 5. NDSM or NHM
b) Divide 3D: Divides the values of two rasters 
on a cell-by-cell basis. The figure 4.6 shows how to B. Removal of Trees
give conditions between two rasters. This can be done Urban trees tend to be confused with
similar between a raster image and a constant. buildings because of similar heights of trees and low-
c) Convert to integer: This step converts each rise buildings. To extract the buildings first the trees
cell value of a raster to an integer by truncation. are removed as shown in figure.

ISSN: 2348 – 8352 www.internationaljournalssrg.org Page 31

 SSRG International Journal of Civil Engineering ( SSRG – IJCE ) – Volume 3 Issue 1 January 2016

Figure 9. Minimum Number of Trees with

Extraction of Buildings

Figure 6. Tree Reduction by Threshold and Parameters

C. Building Extraction
The obtained result confirmed the
assumption that in the image obtained by first
refection point’s interpolation there are more
elements other than buildings that protrude above the Figure 10. 3D Building of a Small Area
terrain, most often these are trees. Whereas in the set
of points from the last reflection, trees are eliminated
to the sizeable extent, last reflection or return is taken VI. CONCLUSION AND FUTURE WORK
shown in following figures In this work, the building extraction has
been completed with certain threshold condition.
Again, further conditions and edge detection
algorithms can be used for better results. The 3D
urban Modelling (built up) is developed for sample
data of Niagara. This concepts is used to develop 3D
for Chennai data is in progress to be analysed in the
work.

REFERENCES
[1] Badea .D, Jocea. A, Negrila A, “3D city modelling by
combination of terrestrial laser scanning data and images,”
Figure 7. Building Extraction Done using ArcGIS University of Civil Engineering, Bucharest, RevCAD –
Shown for Large Area Journal of Geodesy and Cadastre.
[2] G. Hebel. M , Arens. M , Stilla. U, “Utilization of 3d city
models and airborne laser scanning For terrain-based
navigation of helicopters and UAVS,” germany, (IAPRS,
vol. Xxxviii, part 3/w4 - Paris, France, 3-4 september, 2009)
[3] Hongjoo Park, Mahmoud Salah, Samsung Lim, “Accuracy of
3d Models Derived From Aerial Laser Scanning And Aerial
Ortho-Imagery," School of Surveying and Spatial Information
Systems, UNSW Sydney, NSW 2052, Australia.
[4] Jan Bohm, Norbert Haala, “Efficient Integration of Aerial and
Terrestrial laser data for virtual city modeling using
lasermaps,” University of Stuttgart, Institute for
Photogrammetry, Germany; ISPRS WG III/3, III/4, V/3
Workshop "Laser scanning 2005", Enschede, the
Figure 8. Extraction of Building Netherlands.

ISSN: 2348 – 8352 www.internationaljournalssrg.org Page 32

 SSRG International Journal of Civil Engineering ( SSRG – IJCE ) – Volume 3 Issue 1 January 2016

[5] J. Li , *, y. Yuan , m. A. Chapman, “Use of topographic lidar [7] Hongjoo Park, Mahmoud Salah, Samsung Lim, “Accuracy of
point clouds for building Reconstruction based on a gaussian 3d Models Derived From Aerial Laser Scanning And Aerial
markov random field and Markov chain monte carlo Ortho-Imagery", School of Surveying and Spatial
segmentation approach,” Department of geography & Information Systems, UNSW Sydney, NSW 2052, Australia.
environmental management, university of waterloo, 200 PP 304-314
university avenue west, waterloo, ontario n2l 3g1, Canada. [8] Jan Bohm, Norbert Haala, “Efficient Integration of Aerial and
[6] F. Tarsha-Kurdi, T. Landes*, P. Grussenmeyer, E. Smigiel, Terrestrial laser data for virtual city modeling using
“New approach for automatic detection of buildings in lasermaps”, University of Stuttgart, Institute for
airborne Laser scanner data using first echo only,” Photogrammetry, Germany; ISPRS WG III/3, III/4, V/3
Photogrammetry and Geomatics Group, MAP-PAGE UMR Workshop "Laser scanning 2005", Enschede, the
694 - INSA de Strasbourg, 67000 Strasbourg, France, Author Netherlands, September 12-14, 2005,p12
manuscript, Commission III, WG III/3 published in "ISPRS [9] Jurgen Dollner, Thomas H. Kolbe, Falko Liecke, Takis
Comm. III Symposium, Photogrammetric Computer Vision, Sgouros, Karin Teichmann, “The Virtual 3d City Model Of
Bonn:Germany(2006)". Berlin - Managing, Integrating And Communicating
Complex Urban Information”, (2004) PP 34-37

ISSN: 2348 – 8352 www.internationaljournalssrg.org Page 33