department of advanced technology fusion 1 u-science center initial results of a camera-based...

Department of ADVANCED TECHNOLOGY FUSION1u-Science Centeru-Science Center

Initial Results of a Camera-based Traffic Emissions Estimation System for Konkuk

University Complex Testbed

Telescience WG, PRAGMA 14March 2008, Taichung, TW

J.H. Woo, E.I. Kim, S.B. Lim, Y. Sunwoo, K.H. Hong, Y.J. Kim, K.Y. Chung, Y.R. Kim, Y.H. Choi, H Jeong, N.H. Cho

Department of Advance Technology FusionKonkuk University, Seoul, Korea

Department of ADVANCED TECHNOLOGY FUSION2u-Science Center

Research Team

Jung-Hun Woo

- Yoo-Jung Kim - Nam-Ho Cho

Ki-Ho Hong

Eun-Yi Kim

- Ki-Yeong Jeong- Yoon-Hee Shin

Sang-Beom Lim

-You-Lin Jin

iET iIT


Background


U-City is Korea’s vision or strategy for future city and knowledge industry, based on advanced ICT infra-structures

In a “ubiquitous city”, various types of computer/sen-sors and information systems are built into houses, office buildings, streets; make them connected and their services (residential, medical, business, govern-mental) available in a ubiquitous and integrated way

U-Services : Anytime, Anywhere

Need-base Services

Need Testbeds

Background U-City


u-CityInfrastructure

Fire alarm, Seismograph, Flood, Fall down(Penetration, Distortion, Temper-ature, Bright, Vibration,…)

Disaster monitoringFire, Collapsed or Liquidized Soil(Temperature, Smoke, Foun-dation, …)

Physical distribution, Logistic, Quality(Temperature, Humidity, …)

Management of growth, Posi-tion chase, Machine Condition(Position, Tempuature, Humidity …)

Anti-disasterSecurityHealth care &

Medical care

TrackingEnvironmental Risk Management

Others Apps

Healthcare, Emergencycare(Physical condition action,…)

Weather, Water quality, Shop-ping

Agriculture

Air pollution, Vibration, Po-sionous Gas(SOx, NOx, Dust, Distortion, ….)


Konkuk University Complex Our testbed Components : Starcity(Residential &

Commercial), Classic500 (Senior Housing), KU Hospital (KUH), KU Main Campus(KUMC)

Starcity

- Residential : 1310 Condos (58 floors max.)

- Commercial : Multiplex theater, Supermarkets, Mall Area,

Restaurants

Classic500

- Residential : 500 Senior Apartments

- Under Construction KU Hospital

- 870-bed, 83,000 m2, 13 flrs + 4 flrs (un-derground)

- 31 units with 4 major specialized centers

(Outpatient dept) KU Main Campus

- 20,000 students(incl. 2,000 grads), 589 full-time, and

864 part-time faculty. 12 grad schools and 16 colleges

(86 depts and 33 research institutes)

- about 1km2 area (incl. a lake)


Weather Monitoring by AWS

Starcity (Res)

Weather Station (AWS) Elevated Meteorology (60th FL)

Weather Station (AWS) Surface Meteorology


Starcity (Res)

Air Quality Monitoring (NOx, PM10)

AQ Monitoring Station Surface Air Quality

AQ Monitoring Station Elevated Meteorology


Air Quality Monitors


Traffic Monitoring by Webcam

Car : Volume by Types Emissions (MOBILE model)

People : Volume Impact


1st Year Research


1st year research purpose and research flow

Camera based traffic parameters estimation• T

o develop an automatic traffic monitoring system to estimate important traffic parameters, such as velocity & number of vehicles, from video sequences of a one camera

Real time onroad mobile source emissions estimation using measured traffic parameters• E

missions estimation by hour, pollutants, vehicle types

Distributed data management & delivery service• T

o design distributed data management and delivery service for Emissions Monitoring

• To visualize emissions data using short term test data to design and implementation web based user interface & charts


Camera-based Traffic Parameters Estimation


Camera-based Traffic Parameter Estimation

Our goal is to develop an automatic traffic moni-toring system to estimate important traffic pa-rameters from video sequences using only one camera

– Traffic density– Average vehicle velocity– Vehicle type number of vehicles

System Architecture

– Client : Network Camera • Resolution : 320×240, 30 frame/sec• Format : MPEG-4 standard

– Server : PC• Vision system


Camera monitoring install sites

2nd08.01.15

3rd08.01.18~08.01.

21

5th08.02.11~08.2.2

9

4th08.02.11

1st08.01.10~08.01.

12


Camera-based Traffic Parameter Estimation

Three major stages to estimate desired traffic pa-rameters

– Background estimation– Density estimation– Velocity estimation– Number of vehicles estimation

Input Image Stream

Background Model

Image Substraction

Density Estimation

Velocity Estimation

Number of Vehicle

Estimation


Input Video Streaming

L = 3mt t+∆


Background Estimation

The background color is estimated by looking at consecutive frames in the image sequence

– The color value that occurred most often is chosen to be the back-ground value at that point.


ResultsStart Frame End Frame V(m/sec) D N

1 50 99 27.7652 0.24222 2

2 100 149 19.4964 0.23509 1

3 150 199 10.4096 0.22063 1

4 200 249 28.3136 0.31043 3

5 250 299 13.7456 0.17323 1

6 300 349 10.1252 0.2122 1

7 350 399 0 0 0

8 400 449 18.4924 0.21067 1

9 450 499 10.1928 0.17254 1

10 500 549 16.8476 0.21718 1

total 50 549 12.94903 0.199419 12


Realtime onroad mobile source emissions estimation using measured traffic parameters


Emissions Estimation

General emissions equation

• Emissions (ton/year) = No. of registered vehicles(veh.) × VKT(km/day-veh.) × 365day/year × Emission Factor(g/km) × 10-6

• Emissions by pollutants, vehicle types, VKT means vehicle kilometer travelled

Camera-based realtime emissions estimation (this study)

• Emissions(g/h) = VKT by vehicle types (km) × Emission Factor (g/km)• Emissions by hour, pollutants, vehicle types

• Hourly VKT (km) = passing traffic average velocity (km/veh.-h) × No. of vehicles(veh.) × 1hr(h)

Vehicle types in this study

• 7 classifications• Passenger Cars(include taxi), Buses (light duty), Buses(heavy duty), Trucks(compact car,

light duty), Trucks(middle duty, heavy duty, special vehicle), Two Wheelers, RV(Recreational vehicle)


Hourly VKT by vehicle types

TimePassingTraffic

VelocityAvg.

TotalVKT

Unit Veh. km/hr km

2007-02-19-11:00 360.0 44.9 16179.4

2007-02-19-12:00 3253.3 46.8 152235.8

2007-02-19-13:00 266.7 35.9 9567.0

2007-02-19-14:00 386.7 48.4 18709.1

2007-02-19-15:00 226.7 33.3 7550.7

2007-02-20-09:00 186.7 25.0 4657.8

2007-02-20-11:00 346.7 49.5 17145.3

2007-02-20-13:00 346.7 47.3 16387.8

2007-02-20-15:00 333.3 44.8 14937.5

vehicle types No. of registered vehicles (veh.)

Distribution(%)

Passenger Cars(include taxi)

1,873,215 58.9

Buses (light duty) 179,052 5.6

Buses (heavy duty) 15,129 0.5

Trucks(Compact car, light duty)

290,894 9.2

Trucks(middle duty, heavy duty,

special vehicle)89,690 2.8

Two Wheelers 401,844 12.6

RV(Recreational vehicle) 327,992 10.3

Traffic parameter & hourly VKT Distribution of vehicle type

Hourly VKT by distribution of vehicle types

Data from Camera based Traffic Parameter Estimation Data from Seoul registered vehicle in Jan. 2008


Emissions by hour, pollutants, vehicle types

2007-02-19-11:00 　 CO NOx PM10 VOCg/hr g/hr g/hr g/hr

Passenger Cars (include taxi) 7395.1 982.1 247.0 365.9 Buses (light duty) 547.1 302.4 22.4 45.3 Buses (heavy duty) 110.4 354.0 2.4 42.9

Trucks(Compact car, light duty)

811.2 232.9 59.5 39.8

Trucks(middle duty, heavy duty, special vehicle)

1369.6 3456.7 33.4 333.7

Two Wheelers 21482.2 7160.7 　 409.2 RV (Recreational vehicle) 1015.7 217.8 54.4 46.3

Total 32731.4 12706.6 419.1 1283.0

vehicle types CO NOx PM VOCunit g/Km g/Km g/Km g/Km

Passenger Cars(include taxi) 0.7754 0.1030 0.0259 0.0384 Buses (light duty) 0.6002 0.3317 0.0246 0.0497 Buses (heavy duty) 1.4334 4.5958 0.0315 0.5566

Trucks (Compact car, light duty)

0.5477 0.1572 0.0402 0.0269

Trucks(middle duty, heavy duty, special ve-

hicle)2.9992 7.5698 0.0732 0.7309

Two Wheelers 10.5000 3.5000 　 0.2000 RV(Recreational vehicle) 0.6083 0.1304 0.0326 0.0277

Revised emission factor

Hourly VKTby vehicle type

Emissions by hour, pollutants, vehicle types

Data from CAPSS


Distributed data management and service


Our goal is to develop an distributed data management & delivery service in camera base emissions monitoring sys-tem

1st year research goal

– To design distributed data management and delivery service for emis-sions monitoring testbed

– To visualize short term test data in support of design and implementa-tion of web-based user interface

1st year research results

– Distributed data management & delivery service design– Web-based chart and data analysis results through designed UI

Distributed Data Management and Service


Distributed Data Management and Service

Structure DesignCamera data stream convert Module

ASF per-second

25frame JPG

NBPub

NBTopic

K1

K2

K3

KN

Traffic parameter estimation Module

K1NBSub

NBPub

K2NBSub

NBPub

K3NBSub

NBPub

KNNBSub

NBPub

K1 Live track 0

K2 Dead track 0

K3 Busy track 0

KN Dead track 0

Pub Live track 0

DB

U1NBSub

NBPub

U2NBSub

NBPub

U3NBSub

NBPub

UN

NBSub

NBPub

Estimation for air pollutants emis-

sions Module

NBTopic

U1

U2

U3

UN

K1 Live track 0

K2 Dead track 0

K3 Busy track 0

KN Dead track 0

Pub Live track 0

JPG NBTopic

Final Data

process

ManagementManagement

XMLData pair

NBPub

NaradaBrokering publisher

NBSub

NaradaBrokering subscriber

NBTopic

NaradaBrokering Topic


Web-Based Charts of Environmental Data

Http://117.16.146.61:8888/joomla

Analysis charts service

- Emissions Monitoring

- Air Quality Monitoring

- Meteorology Data


PM10 emission VOC emission

The street of Konkuk University front door emissions monitoring test data analysis results : Feb.19 ~20, 2008

Examples of Emission Monitoring


The street of Konkuk University front door air quality monitoring test data analysis results : Feb.19 ~20, 2008

Examples of Air Quality Monitoring

SO2 PM10


ChenHo-Dong(KangDong-Gu, Seoul) meteorology data test analysis results : Feb.11~12, 2008

Examples of Meteorology Data

Wind speed Temperature


Future Improvements

Camera based traffic parameters estimation

Real time onroad mobile source emissions estimation

using measured traffic parameters

Distributed data management & delivery service

Shadow elimination

Vehicle classification Vehicle tracking

Real time emissions estimation by measured vehicle-fuel types Meteorological data monitoring Air quality data monitoring

Automated data transfer implemen-tation Data model, management, and service implementation


Relation to PRAGMA & Workgroup Meetings

Fusion Idea/work evolution

Related Com-ponents

Idea initiation

Pilot implementation Idea expansion Seed project Full-scale projects Related-projects :

Air quality assessment/prediction Sensor-based air quality monitoring Camera-based pollutant emissions monitoring Urban Geography consideration Policy+Science+Technology (Enviro + IT fusion)

: Introductory Fusion Course & Technology Fusion Project (2007. 3~6)

: Techno-Fair, Konkuk Univ (2007. 5): PRAGMA (2007. 9), u-Ecocity Workshop (2007.11): BK21 Internal Project, Konkuk Univ(2007. 11)

: Construction u-City of Seoul by Fu-ture Foresight Analysis(2007. 12)

- National Cyberinfrastructure Project for Construction Research (2007. 9) - Development of artifical intelligence air quality control and management system for subway stations and tunnels (2007. 12)


Constructing u-City of Seoul by Future Foresight Analysis

Building u-City infrastructure for future Seoul

Construction of citizen-friendly city of Seoul from foresight analysis and public needs survey. Devel-opment of nec-essary policies

Ontology develop-ment in support of u-City service

Distributed air mon-itoring and model-ing system

USN-based service development

Sub-project1

Sub-project2

Sub-project3

Plus…

Industrialpartners :

Lukis Metabuild Genosystem SoltluxKosec

SKK Univ. Konkuk Univ. Univ. of Seoul 2.5 million USD for Five Years (2007~2012) from Seoul Met-

ropolitan Government


System Flow Scheme

μ-Gateway

How is air quality around us?

Ask Monitoring Sensor

Air Quality Monitoring Air Quality Modeling Connect with other servies

Sharing Data Respository

Sensor In-formation

Monitor-ing data

Register SensorsCollect monitoring data

Open Data Services Middle-ware

Open data servicesUser-based data services

XML-based Air Quality Data Models

Data Conversion등록

I think it’s okay In this area

StandardizationExpansion

MAMS


Dispersion of Pollutant : CFD Model

Cheonggye Stream

Micro-scale Air Quality Modeling

MAMS (Microscale Air Management System)MAMS (Microscale Air Management System)

3D Building GIS map is being established


대기 환경 모니터링을 위한 시각화 시스템AirScope

• Proceed project and Data Services

• Share Information in do-mestic & Abraod

http://vr.konkuk.ac.kr/subpages/airscope/

department of advanced technology fusion 1 u-science center initial results of a camera-based...

Documents

korea slide

lake slide

integrated way uservices

advanced ict infrastructures

testbeds background

quality temperature

lin jin ietiit slide

water quality