Geo-Based Image Analysis Service In Open Source Cloud Computing Environment

(1)

Free and Open Source Software for Geospatial (FOSS4G)

Conference Proceedings

Volume 15 Seoul, South Korea Article 20

2015

Geo-Based Image Analysis Service In Open Source

Cloud Computing Environment

Sanggoo Kang

Department of Information Systems Engineering, Hansung University

Guseon Yoon

Kiwon Lee

Follow this and additional works at:https://scholarworks.umass.edu/foss4g Part of theGeography Commons

This Paper is brought to you for free and open access by ScholarWorks@UMass Amherst. It has been accepted for inclusion in Free and Open Source Software for Geospatial (FOSS4G) Conference Proceedings by an authorized editor of ScholarWorks@UMass Amherst. For more information, please [email protected].

Recommended Citation

Kang, Sanggoo; Yoon, Guseon; and Lee, Kiwon (2015) "Geo-Based Image Analysis Service In Open Source Cloud Computing Environment," Free and Open Source Software for Geospatial (FOSS4G) Conference Proceedings: Vol. 15 , Article 20.

DOI: https://doi.org/10.7275/R5TX3CKM

(2)

FOSS4G Seoul, South Korea | September 14th – 19th, 2015

GEO-BASED IMAGE ANALYSIS SERVICE IN OPEN SOURCE

CLOUD COMPUTING ENVIRONMENT

Sanggoo Kang1, Guseon Yoon1, and Kiwon Lee2

1

Graduate Student, Department of Information Systems Engineering, Hansung University, Samseon-dong 2-ga, Seongbuk-gu, Seoul, Korea, 136-792

Email: [email protected],[email protected]

2

Professor, Department of Information Systems Engineering, Hansung University, Samseon-dong 2-ga, Seongbuk-gu, Seoul, Korea, 136-792

Email: [email protected]

ABSTRACT

Globally, cloud computing is one of huge trends in ICT communities, exerting its influence on the most information application fields including geo-spatial domain. In general, cloud computing services are being regarded as commercial or public internet data center or infrastructure of computing resources. However, there is no limitation for applications of cloud computing. Especially, it is on the early developing stage in geo-based applications. This project is to present a practical application for geo-based image processing and analysis on open source cloud environment. OpenStack Juno version was applied for open source cloud computing environment. On this cloud environment, PostgreSQL and Django were used as open source for metadata server and web framework, respectively. For geo-based image processing server, OTB was used with GDAL for image manipulating. Some image processing algorithms were implemented in this cloud environment, and provides as a web service for public users. Web-accessible users in this cloud service do not need any software installation and downloading data sets. This full open source approach is expected to be an element geo-spatial information processing linked with cloud computing service.

1. INTRODUCTION

In globally scaled information communication technology (ICT) application, cloud computing leads new paradigm for computing resources uses. The most acceptable one among numerous literal definitions with respect to cloud computing is pay-as-you-go web-accessible computing environment by on-demand typed processing including server hardware, network infrastructure, or software service. Since the early 2010s, cloud computing is widely applied in many application domains, and practical cloud application cases have been reported.In general, cloud-based system shows many advantageous points of economic cost reduction, stability, and scalability (Chou, 2015).

Despite of these huge trend, development of geo-based application in cloud environment is rather slow, it is an early stage. Evangelidiset al. (2014) expected that linkage of cloud computing and standards of Open Geo-spatial Consortium, Inc.(OGC) would expand. Thus it is necessary that experiments for linkage of geo-spatial applications and cloud computing technologies and cloud-based application cases.

The purpose of this work is to demonstrate a cloud-based geo-spatial system which was newly designed and implemented for satellite image processing by open sources covering geo-processing engine and even cloud environments. This system by open source stack and integration supports cross-browser functionalities and does not need any installation processes by users. Computing resources such as data server, web server, data and geo-based image processing engine or client user interface for web mapping are included into

(3)

Geo-based image analysis service in open source cloud computing environment cloud environment, so that end-users with accessing account, basically open accessing, just use web browser. But internal processing in the cloud environment is hidden on the stable mode and quite differ from the non-cloud system.

As the previous works, Lee and Kang (2013) implemented a real-time processing system of satellite image processing for tablet devices by using Amazon elastic compute cloud (EC2) in Amazon Web Service (AWS). Kang and Lee (2014) performed a test case for geo-data fusion application in AWS and OpenStack cloud, one of open source cloud platform. Lee et

al. (2015) developed a prototype for geo-based image analysis system on OpenStack cloud.

This work used more advanced development environments and user interface components, and proposes a new system architecture, different from those previous works.

OpenStack, open source cloud computing platform for public and private clouds, was used for building cloud computing environment, so that generation of instance servers with various performance capability was possible. It means fast adoption and customization of cloud environment according to application purposes.As for OpenStack, it reduce the risks of lock-in associated with proprietary platforms, and offers target-based flexibility and numerous communities and developers choice (OpenStack, 2015). The OpenStack technologies have been built by interrelated projects which control processing, storage, and networking computing resources, capable of being handled by a web-based dashboard or command-line tools.

In many areas of information system business, use cases of OpenStack are increasing every year. But in the geo-spatial field, OpenStack application is a somewhat beginning stage. In spite of current status, integration or linkage of open source cloud computing and geo-based application is a prospective approach. Cloud computing environment helps management and manipulation of large volume of various typed geo data sets and geo-based images including satellite image data.

2. SYSTEM IMPLEMENTATION

The work was fully implemented by open source basis from operating system (OS) to application service level in Table 1. The operating environment is divided intoserverand client. The server was built with OpenStack cloud computing environment and used the operating system, Ubuntu 12.04. Django was used for web framework of Python-based implementation, and it improves security and reliability of fast implemented products. PostgreSQL/PostGIS, open source database management system, was used for metadata cataloging and query process of geo-based image sets. As for image processing engine, Orfeo ToolBox (OTB) providing a variety of open source satellite image processing algorithms was used. In the client side, Bootstrap was used for user interface framework, taking advantage of optimization for multi-screen devices and consistency for displaying contents. As well, JavaScript libraries such as iCheck and selectric based on jQuery were applied for additional user interfacing components. While, OpenLayerswas used for web mapping library, to on-line visualize base map, index image and the processed images.

Figure 1 shows an architecture of the implemented system, with work flow and system modules. Instance server is divided into 4 operating environments. Processing system and user interface system are basically operated on Apache web server and Django. There are many cloud-based user requesting and responding modules. In the geo-based image management module, an important preprocessing for metadata query and display regarding index image archived in the storage units is performed before the image processing or analysis request management module. For actual image processing, OTB and GDAL are used.

(4)

FOSS4G

Among lots of image processing algorithms available from OTB, Image fusion and thresholding scheme are provided for the cloud

attached volume environment means the block storage constructed from the cloud

processing system, and it works storing and managing space of data sets mounted in the instance server. Web client environment is also applied with many open source libraries.

Table 1.

Figure 1. System architecture and components proposed in this work

3. IMPLEMENTATION RESULTS Development Environments Instance Server-side Operating System Web Framework DBMS and Extension Geo-based Image Processing Geo-based Image Manipulation Geo-based Coordinates Manipulation

Client-side

HTML, CSS, and JS Web Mapping Library User Interface Libraries

Among lots of image processing algorithms available from OTB, Image fusion and heme are provided for the cloud-based and real-time processing service. While, attached volume environment means the block storage constructed from the cloud

processing system, and it works storing and managing space of data sets mounted in the nce server. Web client environment is also applied with many open source libraries.

Table 1. Open source list applied in this work

System architecture and components proposed in this work IMPLEMENTATION RESULTS

Development Environments Name / Version

Operating System Ubuntu Server 64bit / 12.04

Framework Django / 1.8.2

and Extension PostgreSQL / 9.3

PostGIS / 2.1

based Image Processing Orfeo ToolBox / 4.4.0

based Image Manipulation GDAL / 1.11.1

based Coordinates Manipulation Proj / 4.8.0 HTML, CSS, and JS Framework

HTML / 5, CSS Bootstrap / 3.3.4 jQuery / 2.1.3

Web Mapping Library OpenLayers / 3.4

User Interface Libraries

iCheck / 1.0.2 selectric / 1.9.3 noUiSlider / 7.0.10

Among lots of image processing algorithms available from OTB, Image fusion and time processing service. While, attached volume environment means the block storage constructed from the cloud-based processing system, and it works storing and managing space of data sets mounted in the

nce server. Web client environment is also applied with many open source libraries.

System architecture and components proposed in this work. Ubuntu Server 64bit / 12.04

PostgreSQL / 9.3 Orfeo ToolBox / 4.4.0 / 3 Bootstrap / 3.3.4 OpenLayers / 3.4 noUiSlider / 7.0.10

(5)

Geo-based image analysis service in open source cloud computing environment Figure 2 represents user interface for end

one page, generic image processing can be carried out by web users: identification index image regarding geo-based image which can be actually processed in the OTB, selection of processing algorithm and concerned parameters choice, and visualization of the final processed result. Figure 2(a) is the menu system for on

(OSM) is the default, and V

provided. Index image is the reduced image by size of the original image data.

processing. Figure 2(b) and (c) are the processing algorithm selection menu and parameter choice menu system from user inputs, respectively. Figure 2(d)

for requesting the cloud-based server processing after several user inputs. In this works, KOrea Multi

Figure 3 shows binary image generation process by thresholding image processing scheme, and Figure 4 isthe produced result by Bayesian image fusion algorithm.

Figure 2. User interface and processing components in the geo

system on OpenStack cloud computing environment

based image analysis service in open source cloud computing environment Figure 2 represents user interface for end-user, accessing cloud-based web system

generic image processing can be carried out by web users: identification index based image which can be actually processed in the OTB, selection of processing algorithm and concerned parameters choice, and visualization of the final cessed result. Figure 2(a) is the menu system for on-line base mapping. Open street map (OSM) is the default, and V-world, open platform for geo-spatial data in Korea, is also

the reduced image by approximately 20% re-sampling

size of the original image data. It is similar to thumbnail image, a term in the digital image processing. Figure 2(b) and (c) are the processing algorithm selection menu and parameter choice menu system from user inputs, respectively. Figure 2(d) is a simple click event button

based server processing after several user inputs.

In this works, KOrea Multi-purpose SATellite image (KOMPSAT 3) sets were used. Figure 3 shows binary image generation process by thresholding image processing scheme,

the produced result by Bayesian image fusion algorithm.

interface and processing components in the geo-based image processing on OpenStack cloud computing environment.

based image analysis service in open source cloud computing environment based web system. In generic image processing can be carried out by web users: identification index based image which can be actually processed in the OTB, selection of processing algorithm and concerned parameters choice, and visualization of the final line base mapping. Open street map spatial data in Korea, is also sampling compared to It is similar to thumbnail image, a term in the digital image processing. Figure 2(b) and (c) are the processing algorithm selection menu and parameter is a simple click event button purpose SATellite image (KOMPSAT 3) sets were used. Figure 3 shows binary image generation process by thresholding image processing scheme,

based image processing .

(6)

FOSS4G

Figure 3.

Figure 4. User interface for

4. CONCLUDING REMARK

Cloud computing and its services are still on the early maturing stage in the geo

data manipulation and processing application. There are many ways and platforms to build a cloud-based application. This study employed OpenStack cloud computing platfo

study based on full open source presents a test case with geo functions. Bayesian image fusion and

algorithm are based on the open source algorithms of OTB. For the further st performance test with multiple source data sets needs for practical uses of mobile cloud in

Figure 3. Binary image generation by thresholding.

User interface for Bayesian image fusion and the processed result ex

CONCLUDING REMARK

Cloud computing and its services are still on the early maturing stage in the geo

data manipulation and processing application. There are many ways and platforms to build a based application. This study employed OpenStack cloud computing platfo

study based on full open source presents a test case with geo-based image processing functions. Bayesian image fusion and image thresholding as geo-based image processing algorithm are based on the open source algorithms of OTB. For the further st performance test with multiple source data sets needs for practical uses of mobile cloud in

y image generation by thresholding.

and the processed result example.

Cloud computing and its services are still on the early maturing stage in the geo-based data manipulation and processing application. There are many ways and platforms to build a based application. This study employed OpenStack cloud computing platform. This based image processing based image processing algorithm are based on the open source algorithms of OTB. For the further study, performance test with multiple source data sets needs for practical uses of mobile cloud in

(7)

Geo-based image analysis service in open source cloud computing environment geo-science application. Also the on-the-fly image matching or image registration on mobile devices for heterogeneous image sets from different sensor are also necessary.

5. ACKNOWLEDGMENTS

This research was partly supported by National Land Space Information Research Program by Ministry of Land, Infrastructure and Transport of Korean government (No. 14NSIP-B080144-01).

6. Reference

Chou, D. C., 2015. Cloud computing: A value creation model, Computer Standards & Interfaces 38, 72-77.

Evangelidis, K., K. Ntouros, S. Makridis, and C. Papatheodorou, 2014. Geospatial services in theCloud, Computers & Geosciences63, 116-122.

Kang, S. and K. Lee, 2014. A Performance Test of Mobile Cloud Service for Bayesian Image Fusion,

Korean Journal of Remote Sensing 30, 445-454.

Lee, K. and S. Kang, 2013. Mobile cloud service of geo-based image processing functions: a test iPad implementation, Remote Sensing Letters 4, 910-919.

Lee, K., S. Kang, T.Chae, and Y. Kim, 2015. Satellite Image Processing Application System onOpenStack Cloud Computing Environment, 2015 ICEO&SI and ICLEI Resilience Forum, Kaohsiung.

OpenStack, 2015.OpenStack User Stories, http://www.openstack.org/user-stories (Accessed July 22, 2015).