• No results found

Traffic Control Model using Image Processing and Cloud Computing based online learning

N/A
N/A
Protected

Academic year: 2021

Share "Traffic Control Model using Image Processing and Cloud Computing based online learning"

Copied!
5
0
0

Loading.... (view fulltext now)

Full text

(1)

www.ijascse.org

Page 28

Traffic Control Model using Image Processing

and Cloud Computing based online learning

Ms. Sonali Rohilla

Dept of Computer Science

Mahamaya Technical University

AbstractRoad congestion due to vehicle traffic is a recurring problem worldwide. In modern life we have to face with many problems one of which is traffic congestion becoming more serious day after day. The daily congestion on today’s roads requires an innovative solution of traffic management systems.

Through this paper, architecture is going to be presented for adapti ve traffic control system with online learning feature. Proposed architecture could provide S ervices such as autonomy, mobility, decision support an d the standard development Environment for traffic management strategies, and so on. Cloud computing allow for an inexpensive use of mass quantities of storage, bandwidth and computing resources using the pay-per-use model on which it thrives. Once information is collected and analyzed further control measures are taken to reduce traffic.

Keywords- Cloud computing, intelligent traffic system, Background extraction, Mobile device sensors, Video Image processing.

I. INT RODUCTION

Most of the tra ffic control signals a re static systems, in which the traffic signal routine is pre-defined and played repetitively. However, real time stories had indicated that such static traffic control signals are non-optimal in various respects [1-4]. The direct outcome of the use of adaptive traffic signal control is the reduction in the vehicle’s waiting time as well as the no-traffic time at the traffic signals [1]. However, besides this direct outcome, other important advantages by using adaptive traffic signal control are reduction in the emissions from vehic les, fuel consumption, congestions (specially on busy routes and in the peak traffic periods), number of stops along corridors, etc. [2-4]. There can be t wo ma jor technologies, which can be us ed for adaptive traffic control systems that are mic rowave range Doppler radar based systems [7] and v ideo based image processing system [1, 5-6, 8-9].

To use Doppler radar based systems is difficu lt for traffic signal control because the system will require mult iple independent radio beams or channels for tracking each vehicle in the tra ffic system and imp le menting such systems may not be feasible ,cu mbersome and e xpensive too.

Ms. Alka Singhal

Dept of Computer Science

Mahamaya Technical Unversity

One of the ma in require ments of the traffic signal control system is that such system should be able to process the data (ca mera images) in rea l t ime. Specified require ment requires that , most adaptive traffic signal control systems should use one basic block of vehic le detection [5-6, 8-9]. So metimes, some additional simple image processing steps are also included. For e xa mp le, simp le background detection and updating is used in [1].

A scheme for lane detection, vehic le trac king and mot ion detection is incorporated in [8]. Though these simple approaches keep the system rea l t ime, the accuracy of such systems with simplistic vehic le detection techniques is comparative ly low [8]. Video images processing requires to deal with various issues and addressing such issues directly improves the performance of vehicle detection algorithms and consequently improve the accuracy of such systems [1, 5-6, 8, 10-12]. The approaches can be image enhancements (contrast, hue correction, equalizat ion, colour balance, etc.), shadow re moval, background detection and updating, noise cancellation /compensation, image correction for illu mination variation (sunny days, cloudy days, nights, etc.), weather variations (snowfall, after snowfall, ra iny day, etc.), and so on.

Another very important issue with the existing adaptive traffic signal control systems is that though there is control over the traffic signal according to the current traffic conditions, the system does not learn fro m its previous e xperiences or neither it update itself for forthcoming conditions that it encounters day by day that are increasing day by day. One way of addressing this issue is to maintain a history of traffic data by storing the data on the local system, send offline support system that collects this data, perform data analysis and machine learn ing and then update the system offline with newly learnt ru les features, and semantics. Thus, the learning and data analysis happens online, not very frequently, and with significant delays after the actual data is gathered. If an online lea rning system should be imp le mented, it shall put heavy demands on the computational require ments of the system and entail significant increase in the processing time. Such adaptive systems are not themselves capable of being scaled so that the whole traffic netwo rk can be made adaptive.

(2)

www.ijascse.org

Page 29

Using GPS was one of the technologies used by a number of researchers to solve the traffic ja m proble m. The [13] work can be considered as an e xa mp le when they have used GPS to detect the on road vehicles speed and control the traffic light to enhance.

II. PROBLEM DEFINIT ION

Increased traffic congestion and associated pollution are forcing everyone in transportation to think about rapid changes in traffic processes and procedures to keep our mobility safe, co mfortable, and economica l.

Though, there has been lot of work done in the area of Traffic Control system through cloud computing. But there were a lways certain limitations with the system. The problems identified with the previous work done are:

1. Controlling the traffic fro m a central location by phasing traffic light fro m a ll a round the city and then sending them to master controller through some commun ication media. The process sometimes results in delays and in turn congestion problems.

2. Using a ca mera a lone to analyze traffic condition was not effic ient enough as it would not work always like in heavy ra ining, sand storm conditions and other unfavorable weather conditions.

3. Using GPS has a limitat ion in terms service reception while driving inside a modern city wh ich a lot of long buildings. Lack of high performance computing platform.

4. Systems were not scalable as they were designed in a centralistic way.

III. PROPOSED MODEL

The system operation is shown as a flowchart in Figure 1. For better coverage of receiv ing traffic informat ion three sources are used video cameras, mobile device detectors and other modes. For avoidance of traditional traffic control system limitat ion the communicat ion network will e xist so that it will be accountable to a series of on-demand service, e xcept that result of the calculations will be send to both, the vehicle and control centers.

The source video camera is capturing the videos from roads and pre-processor unit attached with the system is counting the number of vehicles and sending the data to city traffic center. The model a lso proposes to use mobile device detectors to approximate the number of vehicles in a particular reg ion and other modes like hu man key

interaction can a lso be used to approximate the figure. Modes other than the video camera can be used in situations like poor weather conditions like foggy weather, rain etc.

Figure : 1 Start Other modes Mobile towers Video Processing Unit Physical location detected

Sense traffic data

Country traffic center Intellige nt traffic cloud services Traffic knowledge

Application & Se rvices

Traffic manager

Mobile SMS

Display Boards

Traffic light

controller

CLOUD S ERVER City traffic center Video Frames
(3)

www.ijascse.org

Page 30

Control center packed the requests and send them to wh ich are provider of computational capacity and storage of control center, then the results after p rocessing resend to the center control. Traffic control centers in each city is subsidiaries of the country's traffic control center and informat ion will be send to the database of country traffic center and at the same t ime will archive in database (For more cautiously and according to the needs, the history of the Traffic Control Center are stored).

Each country's traffic control centers are connected together by using the cloud server environment. In this mode, once the sending information fro m c ity's traffic control centers to country's traffic control center this data in cloud application have been updated in environment that cloud make it possible and with using the related graphic supported these data becomes the visual map in graphics form. This map will beco me global as soon as the traffic in formation post will be updated. The so obtained resultant will support all features and route of roads.

Figure 2

Intelligence traffic c loud services will help out in creat ing such maps and will provide required traffic knowledge to control and transmit control signals to the traffic light controller, SMS to the mobile device users to change their route and display boards on roads can also be used to show the maps with informat ion to the commuters to take right decisions to avoid traffic ja ms and time delays in their arrivals.

A. Video Processing Unit

Video processing systems constitutes a stream

processing architecture, in wh ich video fra mes fro m a continuous stream are p rocessed one (or more ) at a t ime. This type of processing is critical in systems that have live video or where the video data is so large that loading the entire set into the workspace is ineffic ient.

There are three types of methods ma inly used in detection of moving object in video processing:

1.Fra me subtraction method[15]

The difference between two consecutive images is taken to determine the presence of moving objects. The calculation in this method is very simple and easy to develop. But in this method it is difficult to obtain a complete outline of moving object; therefore the detection of moving object is not accurate.

2.Optica l Flow Method[14]

Calculat ion of the image optical flow field is done. The clustering processing is done according to the optical flow distribution characteristics of image. Fro m this, the complete move ment information of moving body is found and it detects the moving object from the background. 3.Background

subtraction method[16]

The method in which the difference between the current image and background image is taken for the detection moving objects by using simple a lgorithm. But it is very sensitive to the changes which occur in the external environment and it also has poor anti interference ability.

B. Cloud computing

Cloud compuing is internet-based computing in which la rge groups of re mote servers are networked to allow sharing of data-processing tasks, centralized data storage, and online access to computer services or resources. Clouds can be classified as public, private or hybrid [17].

(4)

www.ijascse.org

Page 31

Figure : 3

Three main types of cloud services:

1. Software as a Se rvice (Saa S) – Th is service provides end-user applications running on a cloud infrastructure that can be accessible fro m various client devices. Exa mples of such applications include accounting, collaboration, customer re lationship manage ment (CRM), enterprise resource planning (ERP), invoic ing, hu man resource management (HRM), content management (CM) and service desk management services, etc.

2. Platform as a Service (PaaS) – This service facilit ies for application design / development, testing, deployment and hosting as well as platform services for team co llaboration, web service integration and ma rshalling, database integration and developer community fac ilitation, etc.

3. Infrastructure as a Service (IaaS) – Th is service provides processing, storage, networks, and other fundamental co mputing resources where the consumers are able to deploy and run their own software. Exa mp les of such services include storage, computation, content delivery network (CDN), service manage ment and etc.

IV. CONCLUSION

Incre ment in tra ffic congestion and the associated pollution problems are forc ing everyone in transportation to think about rapid changes in traffic processes and procedures to keep our mobility safe, co mfo rtable, and economica l. The paper present an extended architecture fo r tra ffic control system that uses image processing, mobile sen sing and online learn ing with the help of cloud computing. A cloud computing based architecture is also proposed which has many advantages over local only system.

Advantages like reduced de mand on co mputation resources, increased accuracy, reduced computation time, and increased fra me rate are reported. Other advantages like scalability and robustness are also discussed. In our opinion, such architecture can be deployed at a very small cost, as the current computation resources at local systems may be sufficient for cloud co mputing architecture.

Further, the system can be easily scaled, maintained, and updated through the cloud servers, providing significant improve ment in the current traffic scenario.

IV. REFERENCES

[1] L. Y. Deng, N. C. Tang, D. L. Lee, C. T.

Wang, and M. C. Lu, "Vision based adaptive

traffic signal control system development," in

International

Conference

on

Advanced

Information Networking and Applications, 2005

,

Los Alamitos, 2005, pp. 385-388.

[2] D. Associates, "Evaluation of an adaptive

traffic signal system," DKS Associates, Oakland,

USA2010.

[3] T. T. Consultants, "Evaluation of main street

adaptive

traffic

signal

system,"

TJKM

Transportation

Consultants,

Pleasanton,

USA2011.

[4] J. M. Hutton, C. D. Bokenkroger, and M. M.

Meyer, "Evaluation of an adaptive traffic signal

system: route 291 in Lee’s summit, Missouri,"

Midwest

Research

Institute

and

Missouri

Department of Transportation, Kansas City,

USA2010.

[5] A. L. P. Douglas, M. Prasad, S. Gowtham, A.

Kalyansundar,

V.

Swaminathan,

and

R.

(5)

www.ijascse.org

Page 32

implementation of real-time stationary vehicle

detection by smart camera at outdoor conditions,"

in

International Conference on Image Processing

,

New York, 2006, pp. 3273-3276.

[6] M. H. Hsiao, H. P. Kuo, H. C. Wu, Y. K.

Chen, and S. Y. Lee, "Object-based video

streaming technique with application to intelligent

transportation systems," in

IEEE International

Conference on Networking, Sensing and Control

,

Taipei, 2004, pp. 315-320.

[7] S. V. Baumgartner and G. Krieger, "Real-time

road traffic monitoring using a fast a priori

knowledge based SAR-GMTI algorithm," in

International Geoscience and Remote Sensing

Symposium

, ed New York, 2010, pp. 1843-1846.

[8] S. Kamijo, T. Kawahara, and M. Sakauchi,

"Vehicle sequence image matching for travel time

measurement

between

intersections,"

in

International Conference on Systems, Man and

Cybernetics

, New York, 2005, pp. 1359-1364.

[9] S. Puntavungkour and R. Shibasaki, "Novel

algorithm of vehicle detection by using new ultra

resolution aerial image, three line scanner," in

International Conference on Systems, Man and

Cybernetics, Vols 1-5, Conference Proceedings

,

New York, 2003, pp. 234-239.

[10] U. Farooq, H. M. Atiq, M. U. Asad, A. Iqbal,

and Z. Azmat, "Design and development of an

image processing based adaptive traffic control

system with GSM interface," in

2nd International

Conference on Machine Vision

, Dubai, 2009, pp.

166-171.

[11] W. F. Lv, L. S. Xu, T. Y. Zhu, B. W. Du, and

D. D. Wu, "An FCD Information Processing

Model under Traffic Signal Control," in

20099th

International Conference on Intelligent Systems

Design andApplications

, New York, 2009, pp.

1161-1166.

[12] Y. Y. Ren, J. F. Xi, and S. Jin, "Study on the

Signal Timing Parameters at Intersections of

Urban Road Under the Condition of Ice-Snow

Weather," in

Second International Conference on

Intelligent

Computation

Technology

and

Automation

, Los Alamitos,

2009, pp. 787-790.

[13] Rahman, M. R., "Method For Controlling

Traffic", USA, IntelCorporation,2002.

[14] Cheng-Ming Huang, Yi-Ru Chenand,

Li-Chen Fu “Real- Time Object Detection and

Tracking on a Moving Camera Platform”

ICROS-SICE International Joint Conference 2009

Fukuoka International Congress Center, Japan

August 18-21, 2009, pp. 717-722.

[15] Lianqiang Niu, Nan Jiang “A Moving

Objects Detection Algorithm Based on Improved

Background

Subtraction”

DOI

10.1109/ISDA.2008.337

978-0-7695-3382-7/08

$25.00 © 2008 IEEE.

[16] Lijing Zhang, Yingli Liang “Motion Human

Detection Based on Background Subtraction”

2010

Second

International

workshop

on

Education

Technology

and

computer

science,2010.

[17]

References

Related documents

In response to this need, this study aims to analyze iso- lation effects according to the difference in the vibration period between the superstructure and the isolation layer in

position and were helpful to study the representation of women and issues of violation of women’s rights in television dramas in Pakistan. As president and an army chief

We then show that although games with positive externalities may not allow any stable imputation on some hierarchical organization, still hierarchies represent the ”most

Urban and Rural professional course students. Hence, the hypothesis which states that “There will be a Significant.. Available online:

In a first step, after definition of topics of interest, and text structure, respectively, author- teams were assigned (i) to further evaluate endpoints in oncology; (ii) to provide

density plot of superposition of F-D curves obtained from unfolding of a single rhodopsin from native ROS plasma membrane, numbers represent the amino acid (aa) number of

In this work we initially expel dimness from picture, and after that enhance the nature of picture and reestablished the deceivability of unique picture and hence

Managers respon- sible for developing international marketing channels where the partners in the strategic alliance have different cultural backgrounds need to be aware of the issue