Image Reconstruction From Scattered Cloud Points...

ISSN: 2278 – 909X International Journal of Advanced Research in Electronics and Communication Engineering (IJARECE)

Volume 6, Issue 11, November 2017

1298

All Rights Reserved © 2017 IJARECE

Abstract— For reconstructing a surface in three dimensions,

point clouds are used. Surface reconstruction helps to recreate

the model of the object. The approximating surface from a

given set of samples is called surface reconstruction. A surface

is constructed by carefully selecting points from unorganized

point clouds in three dimension space and can be converted

into any geometrical shape by removing bad poles. An

algorithm is used for the surface reconstruction from different

point clouds in different geometrical shapes. The intented set of

properties for such algorithms includes: incremental updating,

representation of directional uncertainty, the ability to fill gaps

in the reconstruction, and robustness in the presence of

outliers. Power crust alogorithim is used to remove the bad

poles and filteration is applied. The curve filtering and Point

filtering is applied. Crust algorithm plays an important role

due to its guaranteed quality of mesh generation. Crust

algorithm monitors many parameters of mesh generation and

examines the performance of the algorithm by calculating

parameters. The main aim of the algorithm is to filter out

insignificant data while preserving an acceptable level of

output quality

Index Terms—About four key words or phrases in

alphabetical order, separated by commas.

I. INTRODUCTION

Considering the areas of computer vision and computer

graphics, 3D surface reconstruction is the course of

capturing the shape and appearance of real objects. This

process can be accomplished either by active or passive

methods. If the model is allowed to change its shape in time,

this is referred to as non-rigid construction.

The research of 3D reconstruction has always been a complex

goal to achieve. Using 3D reconstruction one can determine

3D profile of any object, as well as knowing the 3D

coordinate of any point on the profile. The reconstruction of

three-dimensional objects is generally regarded as scientific

problem and core technology for a wide variety of fields, such

as Computer Aided Geometric Design (CAGD), Computer

Graphics, Computer Animation, Computer Vision, medical

imaging, computational science, Virtual Reality, digital

media, etc.

Active Methods

Active methods, i.e. range data methods, given the depth

map, rebuild the 3D profile by numerical approximation

approach and build the object in scenario based on model.

These methods actively interfere with the reconstructed

object.

Passive Method

Passive methods of 3D reconstruction do not interfere with

the reconstructed object; they only use a sensor to measure

the radiance reflected or emitted by the object's surface to

infer its 3D structure through image

understanding. Typically, the sensor is an image sensor in a

camera sensitive to visible light and the input to the method

is a set of digital images (one, two or more) or video.

Surface reconstruction is the method of attaining

three-dimensional complex surface model rapidly and

Image Reconstruction From Scattered Cloud

Points Using Hybrid Filteration

Rumani Sharma*, Arun Bhatia**

*ECE, Kurukshetra University, Haryana, INDIA

**Lecturer of ECE, Kurukshetra University, Haryana, INDIA

https://en.wikipedia.org/wiki/Computer_vision

https://en.wikipedia.org/wiki/Computer_graphics



https://en.wikipedia.org/w/index.php?title=Non-rigid&action=edit&redlink=1

https://en.wikipedia.org/wiki/CAGD

https://en.wikipedia.org/wiki/Computer_Graphics



https://en.wikipedia.org/wiki/Computer_Animation

https://en.wikipedia.org/wiki/Computer_Vision

https://en.wikipedia.org/wiki/Medical_imaging



https://en.wikipedia.org/wiki/Computational_science

https://en.wikipedia.org/wiki/Virtual_Reality

https://en.wikipedia.org/wiki/Digital_media

https://en.wikipedia.org/wiki/Digital_media

https://en.wikipedia.org/wiki/Depth_map



https://en.wikipedia.org/w/index.php?title=Numerical_approximation_approach&action=edit&redlink=1



https://en.wikipedia.org/wiki/Image_understanding



https://en.wikipedia.org/wiki/Sensor

https://en.wikipedia.org/wiki/Digital_images



1299


accurately from three-dimensional statistics collected as a

sample, and it is mostly used in reverse engineering.

Three-dimensional data gathered by measuring device is

generally dense, so it is called Point Cloud data. Point cloud

data are considered as an accumulation of the points in

three-dimensional space, and every point cloud data has

three co-ordinates x, y, z.

According to the various forms of data, point cloud data can

be classified into two forms:

I. Ordered point cloud

II. Scattered point cloud.

There is an approach for the reconstruction of a surface, and

scalar fields defined over it, from scattered data points. The

points are assumed sampled from the surfaces of a 3D object,

and the sampling is assumed to be dense and uniform. Laser

range scanners are capable of producing a dense sampling,

usually organized in a rectangular grid, of an object surface.

Some models allow to measure the RGB components of the

color (i.e. three scalar fields) at each sample point. When the

object has simple shape, this grid of points can be an

acceptable representation. However, objects with a more

complex geometry, e.g. objects with holes, handles, pockets,

cannot be scanned in a single pass, and the various scans are

not easy to merge. Other applications, are like recovering the

shape of a bone from contour data extracted from a CT scan,

requires reconstruction of a surface from data points

arranged in slices. The approach of considering the input

points as unorganized are helpful of generating

cross-derivatives by a uniform treatment of all spatial

directions[20].

In High-quality reconstruction of geometry, a core goal is to

capture detailed (or dense) 3D models of the real scene. Many

systems based on real-time tracking, using sparse maps for

localization rather than reconstruction. Other systems have

used simple point based representations (such as surfels or

aligned point clouds) for reconstruction. Kinect Fusion goes

beyond these point-based representations by reconstructing

surfaces, which more accurately approximate real-world

geometry[14].

In some applications, other information derived from CAD

models, measured values or GPS can also be used and

integrated with the sensor data. . In active and passive

sensors, four other methods for object and scene modelling

can currently be classified :

(1) Image-Based Rendering (IBR): This method does not

consider the generation of a geometric 3D model but, for

specific objects and in view of specific camera motions and

scene conditions, it might be taken as a good technique for

the generation of virtual views. IBR creates novel views of

3D environments directly from input images. The Object

discontinuities, especially in large-scale and geometrically

complex environments, will change the output. Therefore,

the IBR method is generally only used for operation that

require limited visualisation.

(2) Image-Based Modelling (IBM): This is the method

which is widely used for geometric surfaces of architectural

objects or for precise terrain and city modeling. IBM methods

(including photogrammetry) use 2D image measurements

(correspondences) to recover 3D object information from a

mathematical model or they obtain 3D data using methods

such as shape from shading, texture, specularity, contour

(medical applications) and shape from 2D edge gradients.

They are very compact and the sensors are often low cost.

(3) Range-Based Modeling: This method directly occupies

the 3D geometric information of an object. It is based on

costly (at least for now) active sensors and can produce a

highly detailed and accurate representation of most shapes.

The sensors rely on artificial lights or pattern projection. In

the past 25 years many advancements have been made in the

field of solid-state electronics and photonics and many active

3D sensors have been developed. Nowadays many

commercial solutions are available, based on triangulation

(with laser light or border projection), time-of-flight,

continuous wave, interferometry or reflectivity measurement

principles.



1300


(4) Combination Of Image-Based And Range-Based

Modeling: Photogrammetry and laser scanning have been

combined in particular for complex or large architectural

objects, in which no technique by itself can be efficiently and

quickly provide a complete and detailed model. Mostly the

basic shapes such as planar surfaces are determined by

image-based methods while the fine details such as reliefs

employ range sensors[15].

1.5 SURFACE RECONSTRUCION PHASES

Surface Reconstruction phases has the following steps :

Phase 1: Initial Surface Estimation

Phase 2: Mesh Optimization

Phase 3: Smooth Surface Optimization

Phase 1: Initial surface estimation: From an unorganized

set of points, phase 1 creates an initial dense mesh. This

phase evaluates the topology of the surface and generates an

initial estimation of the geometry.

Phase 2: Mesh optimization: Initially the dense mesh

generated in phase 1, phase 2 changes the number of faces by

reducing them and improves the fitting to the data points.

This problem is formulated as optimization of an energy

function that models the trade-off between the competing

goals of accuracy, efficiency and preciseness. The number of

vertices in the mesh, the connections between them, and their

respective positions are taken as free variables in

optimization.

Phase 3: Smooth surface optimization: In third phase, the

surface representation is transformed from a piecewise linear

one (meshes) to a piecewise smooth surface. A new piecewise

smooth representation based on subdivision is now

introduced. These surfaces are perfect for surface

reconstruction because they are simple and easy to

implement, models sharp features precisely and can be fitted

using an extension of the phase 2 optimization[9].

Fig. 1 Surface Reconstruction phases

Various areas considered in Surface Reconstruction

Qualitative Properties

Active Illumination

Calibrated Environment

Uncalibrated Environments

An approach for the reconstruction of a surfaces is

considered, and scalar fields defined over it, from scattered

data points. The points are assumed to be sampled from the

surface of a 3D object, and the sampling is considered to be

dense and uniform.

Laser range scanners are able to produce a dense sampling,

mostly organized in a rectangular grid, of an object surface.

Some models also helps to measure the RGB components of

the color (i.e. three scalar fields) at each sampled point.

When the object has a simple shape, this grid of points can

have a sufficient representation. However, objects with a

more complex geometry, e.g. objects with holes, handles,

pockets, cannot be scanned in a single pass, and the various



1301


scans are not easy to merge. Other applications include

restoring the shape of a bone from contour data extracted

from a CT scan, require reconstruction of a surface from data

points organized in slices. This approach of considering the

input points as unorganized has the advantage of producing

cross-derivatives by the equal treatment of all spatial

directions[20].

II. LITERATURE REVIEW

Fabio Remondino, Sabry El-Hakim (2006) presented a full

pipeline for 3D modelling from terrestrial image data,

regarding the different approaches and analysing all the steps

involved. The main problems and the available solutions are

used for the generation of 3D models from terrestrial images.

Shivali Goel, Rajiv Bansal (2013) developed a system for

image reconstruction from scattered cloud points. Crust

algorithm with umbrella Filtering will be implemented.

Crust algorithm plays a vital role due to its guaranteed

quality of triangular mesh generation. Crust algorithm

monitors the various different parameters of mesh generation

and evaluates the performance of the algorithm by

calculating parameters. The main motive of the algorithm is

to filter out left insignificant data while preserving an

acceptable level of output quality.

Mincheol Yoon et. al. (2007) studied the suitability of

ensembles for surface reconstruction. They experimented

with a largely used normal reconstruction technique and

Multi-level Partitions of Unity accurate for surface

reconstruction, showing that normal and surface ensembles

can be completely combined to handle noisy point sets.

Rajdeep Hooda, Anil Kamboj (2016) studied and analised

various algorithms like crust algorithm, power algorithm and

Delaunay algorithm compared for time taken by the

algorithm for the surface reconstruction.

William Y. Chang (2007) studied techniques for

reconstructing surfaces from points. He describe four main

ideas in the graphics literature: signed distance estimation,

Voronoi-based reconstruction, implicit surface fitting, and

moving least squares surfaces. The main challenges include

reconstruction without surface normals, robustness to noise,

accuracy to sharp features, and provable reconstruction

guarantees.

Rajinder Singh (2015) discussed that virtual machines give

users facility to run different operating system on the current

operating system .With the help of Virtual Machines users

can test the new versions of the software whether they fulfill

the requirements or not. Virtual Machines also help to reduce

the hardware cost of the computer system as one can follow

the desired hardware needs. Main player of Virtual Machines

are Virtua Box, VMware , QEMU, and Windows Virtual PC.

Two Virtual Machine software Virtual Box and VMware are

discussed. Various Features of both the machines are also

discussed.

Bernhard Reitinger et. al. developed a first prototype of a

collective 3D reconstruction system for modeling urban

scenes. An Augmented Reality scout is a person who is

supplied with an ultra-mobile PC, an attached USB camera

and a GPS receiver. The scout is exploring the urban

environment and brings a sequence of 2D images. These

images are explained with GPS data and used iteratively as

input for a 3D reconstruction engine which reconstructs the

3D models on-the-fly. This turns modeling into an

interactive and collaborative task.

Bing Han et. al. presented a different surface fitting

approach for 3D dense reconstruction. They proposed a

non-linear deterministic annealing algorithm to dissolve the

3D sparse structure to separate regions, and estimate the

dense depth map by plane surface fitting. The experimental

results reveal that the new approach can segmented in the 3D

space geometrically and generates smoother dense depth

map.



1302


III. PROPOSED WORK

A. PROBLEM ANALYSIS

Having studied the various previous approaches,

techniques and methods used for the recreation of the surface

in three dimensions revealed out numbers of issues related to

it. The limitations found cannot be eliminated to zero, but

can be reduced to an acceptable level by various iterations

performed by filtering the point clouds.

Following are the various issues related to it:

1) Computational Time: It is the most important

criteria to measure the efficiency of a method. The

technique must be time efficient so as to reduce the

time spent on the computation of the point clouds

and removing unwanted point clouds.

2) Noise: The main factor in all the techniques which

should be reduced to minimum is noise. It appears

in form of distortion of the original surface. Since

accurate model is practically not possible, so an

approximate model should be recreated.

3) Space Utilisation: The true requirement is that the

surface should utilize less space. Hence, the

technique should be chosen such that it minimize

the space utilization by removing the unwanted

point clouds from the model.

4) Minimum Cost: It always remains the prime factor

in every field of technology. High efficiency and low

cost is the main criteria to chosose any technique.

Hence, the technique followed must be yield high

output at low cost.

B. Problem Statement

Some type of filter technique which can remove

insignificant data from original data sets. The

researchers are motivated to reduce the cost of

surface reconstruction by removing insignificant

data, because computation cost is closely related to

the complexity of the data. The main aim of the

algorithm is to filter out insignificant data while

preserving an acceptable level of output quality. In

the previous works, reconstruction time and average

distance had been focused. In addition, there are

many factors which can be concentrated upon.

C. Proposed work

Power crust algorithm is our proposed work. The hybrid

filteration technique is used by the combination of curve

filtering and point filtering. Reduction in space points is

obtained by using Power crust algorithm. This algorithm

gives the better utilization of space. Bad poles are also

removed along with the reduction in space. The filteration

techniques are also used – curve filtering and point filtering

with different values of multiplying factor.

IV. RESULT AND ANAYLSIS

Our aim is to reduce the cost of surface reconstruction by

applying filtering techniques such as curve filtering and

point filtering. Below are shown the results of our proposed

work.

1. Here, we first started with a point cloud image which

is named as ‗hot dogs‘. Initially the size of the

image, without filteration, is 1196826, with

multiplying factor, m =1000. Then we applied curve

filtering on the original image with different values

of curve, C.

Fig 2.. showing original image of hotdog without filteration

a) The reduced size is obtained at 0.2 value of C which is

1155676 and this gives reduction in size of

approximately 3.4%.



1303


Fig. 3 showing hotdog with curve filtering

(threshold), C 0.2, giving around 3.4% reduction in

size.

Table for curve filtering applied on Hotdog and

corresponding reduction in space.

Fig. 4 graph of space reduction in % vs curve

filtering(threshold) for hotdog at various values of curve

filtering.

1. Second iteration we performed on point cloud image

‗knots‘. Without applying filteration, the original

size of the image with a multiplying factor m equal

to 100, is 369935.

Fig. 5 showing original size of ‗knots‘ without

filteration.

a) At C equal to 0.5, the reduced size is 261645, which

gives nearly 29% reduction in size.

Fig. 6 representing curve filtering of 0.5 applied on

knots.

Table of various values of curve filtering applied on

knots and their corresponding %age reduction in

size

Image Name- knot

Original Size-

369935

Curve

Filtering(Threshold)

Space Reduction

in %age

0.3 98

0.4 61

0.5 29

Image Name – hotdogs

Original Size-

1196826

Curve Filtering(Threshold) Space Reduction in

%age

0.1 68

0.2 3.4

0.3 0.000083

0.4 0.23

0.5 0.00033



1304


0.6 0.28

0.7 .00027

Fig. 7 Plot between space reduction and curve

filtering on ‗knots‘.

The above shown table and graph are for various

values curve filtering applied on pts file. There is a

considerable decrease in the size by applying

filteration. The graph plotted between various

values of curve filtering and subsequent decrease in

space obtained.

2. Another filtering is performed on point cloud image

‗bunny‘ original size 5313935 at m 1,00,000.

Fig. 8 showing original image of bunny with

multiplying factor, m 1,00,000

a) At C 0.1 , the size of image is 3510632 which

gives approximately 33% reduction in size.

Fig. 9 curve filtering,C of value 0.1 applied on

bunny.

Table for various values of curve filtering

applied on ‗bunny‘ and corresponding

reduction in size.

Image Name – Bunny

Original Size-

5313935

Curve Filtering(Threshold)

Space Reduction in

%age

0.1 33

0.2 8.24

0.3 2.23

0.4 0.96

0.5 0.36

0.6 0.091

Fig. 10 Graph plotted between space reduction and curve

filtering (threshold)

Above shown is the table and graphical representation of the

curve filteration applied on pts file ,bunny‘. The table shows

that there is considerable amount of decrease in space.

Above shown is the table and graphical representation of the

curve filteration applied on pts file ,bunny‘. The table shows

that there is considerable amount of decrease in space.

Secondly, we will discuss point filtering on point cloud

images knots and hotdog.



1305


1. Our first concern will be ‗hotdog‘ whose original size

is 672748 with multiplying factor, m 100. Now

applying different values of point filtering will yield

different results.

Fig. 11 Original image of ‗hotdog‘ without any

filteration.

a) With point filtering of 0.7, the size of image

obtained is 672487 which gives reduction in

size of 0.038%.

Fig. 12 presenting ‗hotdog‘ with point filtering

of 0.7 applied on it.

Table of point filtering with different values applied on

‗hotdog‘ giving reduction in size.

Image Name- Hotdog Original size- 672748

Point Filtering

(Threshold)

Space Reduction in %age

0.7 0.038

0.8 0.022

0.9 0.002

Fig. 13 Plot between space reduction and threshold values of

point filtering applied on ‗hotdog‘.

Space reduction has been obtained by applying point filtering

on pts file named ‗hot dog‘. The Tabular representation as

well as graphical representation has been shown above and

space reduction is obtained at different values of filtering.

2. Now point filtering is applied on ‗hotdog‘ with

combination of curve filtering, keeping value of

curve filtering constant and varying the value of

point filtering.

Fig. 14 Original image of ‗hot dog‘ without

filteration

At first, the value of curve filtering 0.3 is used and

value of point filtering is varied.

a) For point filtering 0.5, the reduction in size

obtained is 666022 which is 0.99%.



1306


Fig. 15 Curve filtering 0.3 and point filtering 0.5 applied

on ‗hotdog‘ giving 0.99% reduction in size.

a) For 0.6 value of point curve, the image of the

size 665765 is obtained, around 1.03% reduced

size.

Fig. 16 Curve filtering 0.3 and point filtering 0.6 applied on

‗hotdog‘ giving 1.03% reduction in size.

Table Point filtering with different values

applied on ‗hotdog‘ giving reduction in size.

Image Name-

Hotdog

Original size-

672748

Point Filtering

(Threshold)

Space Reduction in

%age

0.5 0.99

0.6 1.03

0.7 1.02

Fig.17 Plot between space reduction and threshold values of

point filtering applied on ‗hotdog‘.

V. CONCLUSION

Here Power Crust algorithm has been implemented with

point filtering and curve filtering. Filtering has been applied

to calculate the bad poles and remove them. Removing the

bad poles from the surface improves the space utilisation and

increases efficiency. These points contain important

information for surface reconstruction. Less geometric points

results in ease of computation hence less computational time.

Geometric patterns become more easier to understand.

VI. FUTURE SCOPE

Planned future work includes improving the performance of

the algorithm. We have already achieved better reductions in

the number of centers but at the cost of slower fitting times.

The main disadvantage is that the image is taken only of pts

extension. This is the limitation with the algorithm so we can

improve its efficiency by dealing with another image format.

Improvement can also be achieved by reducing the noise that

appear in the models so that structure can be more accurately

constructed.

REFERENCES

Agostinho de Medeiros Brito Júnior, Adrião Duarte Dória

Neto, Jorge Dantas de Melo, and Luiz Marcos

Garcia Gonçalves, ―An Adaptive Learning

Approach for 3-D Surface

Reconstruction From Point Clouds‖, IEEE

Transactions On Neural Networks, vol 19, issue 6,

May 28, 2008.

[2] Andrei C. Jalba and Jos B. T. M. Roerdink, Senior

Member, IEEE, ” Efficient Surface Reconstruction

From Noisy Data Using Regularized Membrane

Potentials‖, IEEE Transactions On Image

Processing, vol. 18, no. 5, may 2009.

[3] Vikas Chauhan, Manoj Arora and R. S. Chauhan,

Department of Electronics & Communication

Engg., JMIT Radaur, Yamunanagar, INDIA,

―Comparison of Delaunay Algorithm And Crust

Algorithm for the Optimization of Surface



1307


Reconstruction System‖ , Pelagia Research Library

Advances in Applied Science Research, 2011

[4] Yi-Ling Chen, Student Member, IEEE, and

Shang-Hong Lai, Member, IEEE ,‖ An Orientation

Inference Framework for Surface Reconstruction

From Unorganized Point Clouds‖ , IEEE

Transactions On Image Processing, vol. 20, no. 3,

march 2011.

[5] ZHOU Min, Zhongshan Polytechnic， Zhongshan

， Guangdong,‖ A New Approach of Composite

Surface Reconstruction Based on Reverse

Engineering‖, Elsevier Ltd. Procedia Engineering

00 (2011)

[6] Thomas Schops, Torsten Sattler, Christian Hane,

Marc Pollefeys,ETH Zurich, Switzerland, ‖ 3D

Modeling on the Go: Interactive 3D Reconstruction

of Large-Scale Scenes on Mobile Devices‖ , 3DV

2015, ©2015 IEEE.

[7] Ilya Braude, Jeffrey Marker, David Breen,

Department of Computer Science, Drexel

University, Philadelphia, PA, Ken Museth,

Department of Science and Technology, Link¨oping

University, Norrk¨oping, Sweden, Jonathan

Nissanov, Department of Neurobiology & Anatomy,

Drexel University College of Medicine,

Philadelphia, PA, ―Contour-Based Surface

Reconstruction using MPU Implicit Models‖ ,

Elsevier Science.

[8] David Page, Andreas Koschan, and Mongi Abidi,

University of Tennessee, USA,‖ Methodologies and

Techniques for Reverse Engineering–The Potential

for Automation with 3-D Laser Scanners‖.

[9] Bruno Dutailly, Helene Coqueugniot, Pascal

Desbarats, Stefka Gueorguieva, Remi Synave,

Universite Bordeaux, CNRS, Talence Cedex,

France, ―3d Surface Reconstruction Using HMH

Algorithm‖, IEEE, International Conference on

Image Processing December 2009.

[10] Nidhi Sharma, CSE & Kurukshetra University,

Haryana, India, ‖ A Survey on Surface

Reconstruction Algorithms for 3-D Unorganized

points‖, International Journal of Advanced

Research in Computer Science and Software

Engineering, Volume 4, Issue 4, April 2014.

[11] Bruno Dutailly, Helene Coqueugniot, Universite

Bordeaux, CNRS, UMR 5199 PACEA, F-33405,

Talence Cedex, France, Pascal Desbarats, Stefka

Gueorguieva, Remi Synave, Universite Bordeaux,

CNRS, UMR 5800 LaBRI, F-33405, Talence

Cedex, France, ―3D Surface Reconstruction Using

Hmh Algorithm‖, International Conference on

Image Processing · December 2009 DOI:

10.1109/ICIP.2009.5413911, Source: IEEE Xplore.

[12] Bernhard Reitinger, Dieter Schmalstieg, Institute

for Computer Graphics and Vision, Graz University

of Technology, Austria, Christopher Zach, VRVis

Research Center, Austria, ―Augmented Reality

Scouting for Interactive 3D Reconstruction‖, IEEE,

2007.

[13] Hailin Jin, Anthony J. Yezzi, Yen-Hsi Tsai,

Li-Tien Cheng, and Stefano Soatto , ―Estimation of

3D Surface Shape and Smooth Radiance from 2D

Images: A Level Set Approach‖, Journal of

Scientific Computing, Vol. 19, Nos. 1–3, December

2003, Received May 3, 2002; accepted (in revised

form) September 14, 2002.

[14] Shahram Izadi, Otmar Hilliges, Pushmeet Kohli,

Jamie Shotton, Steve Hodges, Microsoft Research

Cambridge, UK, Dustin Freeman, University of

Toronto, Canada, Andrew Fitzgibbon, David Kim,

Newcastle University, UK, David Molyneaux,

Lancaster University, UK, Richard Newcombe,

Andrew Davison, Imperial College London, UK,

―KinectFusion: Real-time 3D Reconstruction and

Interaction Using a Moving Depth Camera‖,

October 16, 2011, research conducted at Microsoft

Research Cambridge, UK.

[15] Fabio Remondino , Swiss Federal Institute of

Technology (ETH), Zurich, Sabry El, National

Research Council, Ottawa, Canada, ―Image-Based



1308


3d Modelling: A Review‖, The Photogrammetric

Record 21(115): 269–291 (September 2006).

[16] SHIVALI GOEL , CSE, Kurukshetra University,

Kurukshetra, India, RAJIV BANSAL, Assistant

Professor (CSE), JMIT, Yamuna Nagar, Haryana,

India, ‖ Implementation Of Surface Reconstruction

Using Scattered Point Cloud With Crust

Algorithm‖, International Journal of Computer

Science and Engineering (IJCSE), ISSN

2278-9960,Vol. 2, Issue 4, Sep 2013, 185-190 ©

IASET.

[17] Jing Tong, Jin Zhou, Ligang Liu, Zhigeng Pan, and

Hao Yan, ‖ Scanning 3D Full Human Bodies using

Kinects‖.

[18] Shivali Goel ,CSE, Kurukshetra University,India,

Rajiv Bansal, Asst Prof. (CSE), JMIT ,India, ―

Surface Reconstruction Using Scattered Cloud

Points ―International Journal of Advanced Research

in Computer Science and Software Engineering,

Volume 3, Issue 5, May 2013.

[19] J. C. Carr, T. J. Mitchell, Applied Research

Associates NZ Ltd, University of Canterbury, J. B.

Cherrie, W.R. Fright, B. C. McCallum, Applied

Research Associates NZ Ltd, R. K. Beatson,

University of Canterbury, ―Reconstruction and

Representation of 3D Objects with Radial Basis

Functions‖, 2001.

[20] Chandrajit L. Bajaj, Fausto Bernardini, Guoliang

Xu, Department of Computer Sciences, Purdue

University, ―Automatic Reconstruction of Surfaces

and Scalar Fields from 3D Scans‖, Kluwer

Academic publisher by Norwell, MA, 2004.

[21] Mincheol Yoon, Yunjin Lee, Seungyong Lee,

Department of Computer Science and Engineering,

POSTECH, San 31, Hyoja-dong, Pohang, 790-784,

Republic of Korea, Ioannis Ivrissimtzis,

Department of Computer Science, Durham

University, South Road, Durham DH1 3LE, UK,

Hans-Peter Seidel, Max-Planck-Institut f¨ur

Informatik, Stuhlsatzenhausweg 85, 66123

Saarbrucken, Germany, ‖ Surface and normal

ensembles for surface reconstruction‖

Sciencedirect, Computer-Aided Design 39 (2007)

408–420, Received 14 September 2006; accepted 16

February 2007.

[22] William E. Lorensen, Harvey E. Cline, General

Electric Company, Corporate Research and

Development, Schenectady, New York 12301,

―Marching Cubes: A High Resolution 3D Surface

Construction Algorithm‖, Computer Graphics,

Volume 21, Number 4,July 1987. William E.

[23] Brian Curless and Marc Levoy, Stanford University,

―A Volumetric Method for Building Complex

Models from Range Images‖, SIGGRAPH '96

Proceedings of the 23rd annual conference on

Computer graphics and interactive techniques,

1996.

[24] Amin Alqudah,Computer Engineering

Department, Hijjawi Faculty for Engineering

Technology, Yarmouk University, Irbid, Jordan,

―Survey of Surface Reconstruction Algorithms‖,

Journal of Signal and Information Processing,

2014, 5, 63-79, Published Online August 2014 in

SciRes, Received 8 April 2013; revised 15 May

2013; accepted 7 June 2013.

[25] Gang Zeng, Long Quan, Dep. of Computer Science,

HKUST, Clear Water Bay, Kowloon, Hong Kong,

Sylvain Paris, GRAVIR-IMAG, INRIA

Rhone-Alpes, 38334, Saint Ismier, France, Maxime

Lhuillier, LASMEA, UMR CNRS 6602, Universite

Blaise-Pascal, 63177, Aubiere, France, ―Surface

Reconstruction by Propagating 3D Stereo Data in

Multiple 2D Images‖, Springer-Verlag Berlin

Heidelberg 2004.

[26] Rajdeep Hooda, Anil kamboj, ―A Survey of Surface

Reconstruction‖, International Journal of Advanced

Research in Electronics and Communication

Engineering (IJARECE) Volume 5, Issue 4, April

2016.



1309


[27] William Y. Chang, Department of Computer

Science and Engineering, University of California,

San Diego, ‖ Surface Reconstruction from Points‖

[28] Andrei C. Jalba and Jos B. T. M. Roerdink Senior

Member, IEEE, ―Efficient Surface Reconstruction

using Generalized Coulomb Potentials‖, IEEE

Transactions On Visualization And Computer

Graphics, vol. 13, no. 6, november/december 2007.

[29] Rajinder Singh, Dept. of Computer Science and

Applications, Panjab University S.S.G. R.C.

Hoshiarpur Punjab, India, ― A Comparison of

Features of VirtualBox and VMware‖, International

Journal of Advanced Research in Computer Science

and Software Engineering, Volume 5, Issue 9,

September 2015.

[30] V. Estellers, M.A. Scott, S. Soatto, ―Robust Surface

Reconstruction‖Society for Industrial and Applied

Mathematics, Vol. xx, pp. x x–x

[31]

https://en.wikipedia.org/wiki/Ubuntu_(operating_

system)

http://blog.sudobits.com/2012/07/20/top-17-termin

al-commands-every-ubuntu-user-should-know-abo

ut/

[32] http://www.geomview.org/

[33]

https://www-uxsup.csx.cam.ac.uk/pub/doc/suse/su

se9.0/userguide-9.0/ch24s04.html.

[34] https://help.ubuntu.com/community/VirtualBox

[35]

https://en.wikipedia.org/wiki/GNU_Compiler_Co

llection

http://blog.sudobits.com/2012/07/20/top-17-terminal-commands-every-ubuntu-user-should-know-about/



http://www.geomview.org/

https://www-uxsup.csx.cam.ac.uk/pub/doc/suse/suse9.0/userguide-9.0/ch24s04.html

https://www-uxsup.csx.cam.ac.uk/pub/doc/suse/suse9.0/userguide-9.0/ch24s04.html

https://help.ubuntu.com/community/VirtualBox

https://en.wikipedia.org/wiki/GNU_Compiler_Collection

https://en.wikipedia.org/wiki/GNU_Compiler_Collection

Image Reconstruction From Scattered Cloud Points...

Documents

Transcript of Image Reconstruction From Scattered Cloud Points...