A SEMINAR REPORT ON

SPEED DETECTION OF MOVING VEHICLES (USING TRAFFIC ENFORCEMENT CAMERAS)

BY

AKPEOKHAI EMMANUEL OSHOGWE

1

CHAPTER ONE

INTRODUCTION

1.1 Background of the study

Although there is good road safety performance, yet the numbers of people

killed and injured on our roads remain unacceptably high. So the roads safety

strategy was introduced to support the new casualty reduction target. There are

many different factors that lead to traffic collisions and casualties. The main

reason is excess speed of vehicles. We use traffic lights and other traffic

managers to reduce the speed. One among them is speed cameras.

Speed cameras on the side of urban and rural roads are usually placed to catch

transgressors of the stipulated speed limit for that road. Laws are passed making

excessive speed an offence. The speed cameras are used to identify those drivers

that pass by them when they exceed the stipulated speed limit. At first glance

this seemed to be reasonable that the road users do not exceed the speed limit

and this is a good idea because it increases road safety, reduces accidents and

protect other road users and pedestrians.

The police can't be everywhere to enforce the speed limit, so the cameras are to

do this work on any one who's got an ounce of Commons sense, so almost

everyone slowdown for the speed Camera. Hear we finally have a solution to

the speeding problem. Now if we are to assume that speed cameras are the only

way to make driver's slowdown, and they work efficiently, then we would

expect a great number of these everywhere and that they would be highly visible

and identifiable to make a driver slow down.

The system automatically captures image of a moving vehicle and records the

data parameters, such as date, time, speed operator and location, etc. A capture

2

window that comprises a predetermined range of distances of the system from

the moving vehicle can be set by the operator so that the image of the moving

vehicle is automatically captured when it enters the capture window. The

capture window distance can be entered manually through a keyboard or

automatically using the laser speed gun. Automatic focusing is provided using

distance information from the laser speed gun. Traffic management and

information systems rely on a suite of sensors for estimating traffic parameters.

Magnetic loop detectors are often used to count vehicles passing over them.

Vision­based video monitoring systems offer a number of advantages. In

addition to vehicle counts, a much larger set of traffic parameters such as

vehicle classifications, lane changes, etc. can be measured. Besides, cameras are

much less disruptive to install than loop detectors. Vehicle classification is

important in the computation of the percentages of vehicle classes that use

state­aid streets and highways. The current situation is described by outdated

data and often, human operators manually count vehicles at a specific street.

The use of an automated system can lead to accurate design of pavements (e.g.,

the decision about thickness) with obvious results in cost and quality. Even in

large metropolitan areas, there is a need for data about vehicle classes that use a

particular street. A classification system can provide important data for a

particular design scenario. Here system uses a single camera mounted on a pole

or other tall structure, looking down on the traffic scene. It can be used for

detecting and classifying vehicles in multiple lanes and for any direction of

traffic flow. The system requires only the camera calibration parameters and

direction of traffic for initialization.

3

CHAPTER 2

TRAFFIC ENFORCEMENT CAMERA

2.1 WHAT IS A TRAFFIC ENFORCEMENT CAMERA

A traffic enforcement camera (also red light camera, road safety camera, road

rule camera, photo radar, photo enforcement, speed camera, Gatso) is an

automated ticketing machine. It may include a camera which may be mounted

beside or over a road or installed in an enforcement vehicle to detect traffic

regulation violations, including excess speed, vehicles going through a red

traffic light, unauthorized use of a bus lane, for recording vehicles inside a

congestion charge area.

Fig 1: Gasto speed camera

The latest automatic number plate recognition systems (ANPR) can be used for

the detection of average speeds and raise concerns over loss of privacy and the

potential for governments to establish mass surveillance of vehicle movements

and therefore by association also the movement of the vehicle's owner. Vehicles

owners are often required by law to identify the driver of the vehicle and a case

4

was taken to the European Court of Human Rights who found that the Human

Rights Act 1998 was not being breached. Some groups, such as the National

Motorists Association in the USA, claim those systems "encourages

revenue­driven enforcement" rather than the declared objectives.

2.2 HISTORY

The concept of the speed camera can be dated back to at least 1905; Popular

Mechanics reports on a patent for a "Time Recording Camera for Trapping

Motorists" that enabled the operator to take time­stamped images of a vehicle

moving across the start and endpoints of a measured section of road. The

timestamps enabled the speed to be calculated, and the photo enabled

identification of the driver.

The Dutch company Gatsometer BV, which was founded in 1958 by rally driver

Maurice Gatsonides, produced the 'Gatsometer'. Gatsonides wished to better

monitor his average speed on a race track and invented the device in order to

improve his lap times. The company later started supplying these devices as

police speed enforcement tools. The first systems introduced in the late 1960s

used film cameras to take their pictures. Gatsometer introduced the first red

light camera in 1965, the first radar for use with road traffic in 1971 and the first

mobile speed traffic camera in 1982.

From the late 1990s, digital cameras began to be introduced. Digital cameras

can be fitted with a network connection to transfer images to a central

processing location automatically, so they have advantages over film cameras in

speed of issuing fines, maintenance and operational monitoring. However,

5

film­based systems may provide superior image quality in the variety of lighting

conditions encountered on roads, and are required by courts in some

jurisdictions. New film­based systems are still being sold, but digital pictures

are providing greater versatility and lower maintenance and are now more

popular with law enforcement agencies.

Fig 2: Older traffic enforcement cameras

2.3 TRAFFIC ENFORCEMENT CAMERAS

2.3.1 Fixed­speed and red light cameras

With the introduction of digital technology, it is becoming more common for

red­light cameras to also function as fixed speed cameras. Most red­light

cameras and many speed cameras are fixed­site systems mounted in boxes or on

poles beside the road. They are also often attached to gantries over the road, or

to overpasses or bridges. In some areas such as New South Wales in Australia,

there are more pre­configured fixed camera sites than actual cameras, with the

6

camera equipment being rotated periodically between the sites. The system

continuously monitors the traffic signal and the camera is triggered by any

vehicle entering the intersection above a preset minimum speed and following a

specified time after the signal has turned red.

Fig 3: Fixed­speed and red light cameras

Fixed camera systems can be mounted in boxes or on poles beside the road or

attached to gantries over the road, or bridges. Cameras can be concealed or

dazzled.

2.3.2 Mobile Speed Cameras

Mobile speed cameras may be hand­held, tripod mounted, or vehicle­mounted.

In vehicle­mounted systems, detection equipment and cameras can be mounted

to the vehicle itself, or simply tripod mounted inside the vehicle and deployed

out a window or door. If the camera is fixed to the vehicle, the enforcement

vehicle does not necessarily have to be stationary, and can be moved either with

or against the flow of traffic. In the latter case, depending on the direction of

7

travel, the target vehicle's relative speed is either added or subtracted from the

enforcement vehicle's own speed to obtain its actual speed. The speedometer of

the camera vehicle needs to be accurately calibrated.

Fig 4: Tripod mounted mobile speed camera

2.3.3 Bus lane enforcement cameras

Some bus lane enforcement cameras use a sensor in the road which triggers a

number plate recognition camera which compares the vehicle registration plate

with a list of approved vehicles and records images of other vehicles. Other

systems use a camera mounted on the bus, for example in London where they

monitor Red routes on which stopping is not allowed for any purpose (other

than taxis and disabled parking permit holders).

On Monday, February 23, 2009, New York City announced testing camera

enforcement of bus lanes on 34th Street in Midtown Manhattan where a New

York City taxi illegally using the bus lanes would face a fine of $150

adjudicated by the New York City Taxi and Limousine Commission.

8

2.2.4 Stop sign enforcement cameras

In 2007, the Mountains Recreation and Conservation Authority (MRCA), in

California, installed the first stop sign cameras in the United States. The five

cameras are located in state parks such as Franklin Canyon Park and Temescal

Gateway Park. The operator, Redflex Traffic Systems Inc., is paid $20 per

ticket. The fine listed on the citation is $100. In 2010 a class action suit was

filed against MRCA.

9

CHAPTER 3

CONTRIBUTION/PROPOSAL

3.1 TRAFFIC ENFORCEMENT CAMERA SYSTEM

3.1.1 AVERAGE SPEED CAMERAS SYSTEM

These systems operate using automatic digital technology. Cameras are

mounted on columns at the side of the road. By placing the cameras at certain

points, the speed of vehicles can be monitored throughout the traffic

management area. Each pair of cameras consists of two digital video cameras,

linked by cable or radio wave. The cameras act as a speed controlled zone with

groups of cameras linked to create a speed controlled network.

The video cameras continuously capture images of vehicles. The number plates

are read using Automatic Number Plate Recognition (ANPR) and the average

speed of the vehicle is calculated between the two cameras. If this exceeds the

speed limit, an offence record is created.

Fig 5: Average speed

camera system

10

How average speed camera systems works

Average speed camera systems work by calculating the speed of a vehicle over

a distance. A vehicle enters the gateway and the camera using Automatic

Number Plate Recognition (ANPR) records the number plate. When the vehicle

passes through the exit gateway the camera matches the number plate and

carries out a simple time over distance calculation and if the vehicle has been

travelling above the speed limit the offence is recorded. If there is no offence,

the camera does not retain details of the vehicle number plates. Below is a

diagram illustrating how an average speed camera system works.

Fig 6: how the average speed camera works

Why we use an average speed camera system

Average speed cameras help to make roads safer by encouraging drivers

to maintain a consistent speed limit.

11

Average Speed Cameras are one example of new Intelligent Transport

Systems (ITS). The information collected by the cameras is used to make

our roads safer.

Roadwork’s can prove a dangerous working environment for contractors,

with drivers having to react to contra flows, narrow lanes and changes in

road layout. The operational safety of the site is enhanced when speed

limits are reduced.

The average speed camera system has been installed to ensure

compliance with the reduced temporary 40mph speed limit. The speed

limit has been reduced for the safety of the construction worker and all

road users. A positive effect of average speed cameras at road works is

that traffic is known to flow smoothly.

3.1.2 ANPR

The Automatic Number Plate Recognition (ANPR) is a fixed or mobile speed

camera system that measure the time taken by a vehicle to travel between two or

more fairly distant sites (from several hundred meters to several hundred

kilometers apart). These cameras time vehicles over a known fixed distance, and

then calculate the vehicle's average speed for the journey. The name derives

from the fact that the technology uses infrared cameras linked to a computer to

"read" a vehicle's registration number and identify it in real­time.

In principle, it is not possible (as in the case of a single speed camera) to slow

down momentarily while passing one of the cameras in order to avoid

12

prosecution, as the average speed over a distance rather than the instantaneous

speed at a single point is calculated.

In the case of the Australian SAFE­T­CAM system, ANPR technology is also

used to monitor long distance truck drivers to detect avoidance of legally

prescribed driver rest periods. The state of Victoria has recently introduced an

ANPR system for monitoring passenger vehicles.

In the United Kingdom, automatic number plate recognition (ANPR)

average­speed camera systems are known by the Home Office as SVDD (Speed

Violation Detection Deterrent). More commonly, they are known by the public

by their brand name ­ SPECS (Speed Enforcement Camera System), a product

of Speed Check Services Limited, or just as speed cameras/traps. They are

frequently deployed at temporary roadwork sites on motorways, and are

increasingly being used at fixed positions across the UK.

Automatic number plate recognition systems can also be used for multiple

purposes, including identifying untaxed and uninsured vehicles, stolen cars and

potentially mass surveillance of motorists.

ANPR is sometimes known by various other terms:

o Automatic license plate recognition (ALPR)

o Automatic vehicle identification (AVI)

o Car plate recognition (CPR)

o License plate recognition (LPR).

13

Development history

The ANPR was invented in 1976 at the Police Scientific Development Branch

in the UK. Prototype systems were working by 1979, and contracts were let to

produce industrial systems, first at EMI Electronics, and then at Computer

Recognition Systems (CRS) in Wokingham, UK. Early trial systems were

deployed on the A1 road and at the Dart ford Tunnel.

Components

The software aspect of the system runs on standard home computer hardware

and can be linked to other applications or databases. It first uses a series of

image manipulation techniques to detect, normalize and enhance the image of

the number plate, and then optical character recognition (OCR) to extract the

alphanumeric of the license plate. ANPR systems are generally deployed in one

of two basic approaches: one allows for the entire process to be performed at the

lane location in real­time, and the other transmits all the images from many

lanes to a remote computer location and performs the OCR process there at

some later point in time. When done at the lane site, the information of the plate

alphanumeric, date­time, lane identification, and any other information that is

required is completed in somewhere around 250 milliseconds.

14

Fig 7: Flow chat of the ANPR

Technology used in ANPR

ANPR uses optical character recognition (OCR) on images taken by cameras.

When Dutch vehicle registration plates switched to a different style in 2002, one

of the changes made was to the font, introducing small gaps in some letters

(such as P and R) to make them more distinct and therefore more legible to such

systems, as shown below. Some license plate arrangements use variations in

font sizes and positioning. ANPR systems must be able to cope with such

differences in order to be truly effective. More complicated systems can cope

with international variants, though many programs are individually tailored to

each country.

15

Fig 8: some license plate

The cameras used can include existing road­rule enforcement or closed­circuit

television cameras, as well as mobile units, which are usually attached to

vehicles. Some systems use infrared cameras to take a clearer image of the

plates

Installing ANPR Cameras on Law Enforcement Vehicles

Installing ANPR cameras on law enforcement vehicles requires careful

consideration of the juxtaposition of the cameras to the license plates they are to

read. Using the right number of cameras and positioning them accurately for

optimal results can prove challenging, given the various missions and

environments at hand. Highway patrol requires forward­looking cameras that

span multiple lanes and are able to read license plates at very high speeds. City

patrol needs shorter range, lower focal length cameras for capturing plates on

parked cars. Parking lots with perpendicularly parked cars often require a

specialized camera with a very short focal length. Most technically advanced

systems are flexible and can be configured with a number of cameras ranging

from one to four which can easily be repositioned as needed. States with

16

rear­only license plates have an additional challenge since a forward­looking

camera is ineffective with incoming traffic. In this case one camera may be

turned backwards

Fig 9: ANPR cameras on law enforcement vehicles

Recent advances in technology have taken automatic number plate recognition

(ANPR) systems from fixed applications to mobile ones. Scaled­down

components at more cost­effective price points have led to a record number of

deployments by law enforcement agencies around the world. Smaller cameras

with the ability to read license plates at high speeds, along with smaller, more

durable processors that fit in the trunks of police vehicles, allow law

enforcement officers to patrol daily with the benefit of license plate reading in

real time, when they can interdict immediately.

Despite their effectiveness, there are noteworthy challenges related with mobile

ANPRs. One of the biggest is that the processor and the cameras must work fast

17

enough to accommodate relative speeds of more than 100 mph (160 km/h), a

likely scenario in the case of oncoming traffic. This equipment must also be

very efficient since the power source is the vehicle battery, and equipment must

be small to minimize the space it requires. Relative speed is only one issue that

affects the camera's ability to actually read a license plate. Algorithms must be

able to compensate for all the variables that can affect the ANPR's ability to

produce an accurate read, such as time of day, weather and angles between the

cameras and the license plates. A system's illumination wavelengths can also

have a direct impact on the resolution and accuracy of a read in these

conditions.

3.1.3 SPECS CAMERAS SYSTEM

SPECS is an average speed measuring speed camera system manufactured by

the Speed Check Services Limited, from which it takes its name (Speed Check

Services). They are one of the systems used for speed limit enforcement in the

United Kingdom.

About Specs Cameras

SPECS cameras operate as sets of two or more cameras installed along a fixed

route that can be from 200 meters (660 feet) to 10 kilometers (6.2 mi) in length.

They work by using an automatic number plate recognition (ANPR) system to

record a vehicle's front number plate at each fixed camera site. As the distance

is known between these sites, the average speed can be calculated by dividing

18

this by the time taken to travel between two points. The cameras use infrared

photography, allowing them to operate both day and night.

There is a popular misconception that the Home Office has approved the

SPECS system for single­lane use only. According to this theory, a motorist can

therefore switch lanes between cameras and claim non­approval to avoid

prosecution for speeding. However the marketing director of the manufacturer,

Speed checks Services Ltd, has stated that this theory is "categorically untrue".

3.2 LEGAL ISSUES

There are a number of legal issues which arise as a result depending on local

laws and the procedures used by the enforcing bodies. Various legal issues arise

from such cameras and the laws involved in how cameras can be placed and

what evidence is necessary to prosecute a driver varies considerably in different

legal systems.

One issue is the potential conflict of interest when private contractors are paid a

commission based on the number of tickets they are able to issue. Pictures from

the San Diego red light camera systems were ruled inadmissible as court

evidence in September 2001. The judge said that the "total lack of oversight"

and "method of compensation" made evidence from the cameras "so

untrustworthy and unreliable that it should not be admitted".

Some U.S. states and provinces of Canada such as Alberta operate "owner

liability" where it is the registered owner of the vehicle who must pay all such

fines regardless of whether he was driving at the time of the offense, although

19

they do release the owner from liability if he signs a form identifying the actual

driver and that individual pays the fine. These states do not issue demerit points

for camera infractions, which has been criticized by some as giving a "license to

speed" to those who can more easily afford speeding fines.

In a few U.S. states (including California), the cameras are set up to get a "face

photo" of the driver;. This has been done because in those states, red light

camera tickets are criminal violations, and criminal charges must always name

the actual violator. In California, that need to identify the actual violator has led

to the creation of a unique investigatory tool, the fake "ticket."

3.3 Surveillance

Police and government have been accused of "Big Brother tactics" in

over­monitoring of public roads, and of "revenue raising" in applying cameras

in deceptive ways to increase government revenue rather than improve road

safety.

3.4 Revenue not safety

In 2010 a campaign was set up against a speed camera on a dual carriageway in

Poole, Dorset in a 30 mph area in the United Kingdom. Of which had generated

£1.3m of fines every year since 1999. The initial Freedom of information

request was refused and the information was only released after an appeal to the

Information Commissioner.

In May 2010 the new Coalition government said that the Labor’s 13­year war

on the motorist is over' and that the new government 'pledged to scrap public

funding for speed cameras. In July Mike Penning, the Road safety minister

20

reduced the Road Safety Grant for the current year to Local Authorities from

£95 million to £57 million saying that local authorities had relied too heavily on

safety cameras for far too long and that he was pleased that some councils were

now focusing on other road safety measures. It is estimated that the as a result

the Treasury is now distributing £40 million less in Road Safety Grant than is

raised from fines in the year. Dorset and Essex announced plans to review

camera provision with a view to possibly ending the scheme in their counties;

however Dorset strongly affirmed its support for the scheme, albeit reducing

financial contributions in line with the reduction in government grant. Seven

counties also announced plans to turn off some or all of their cameras, amidst

warnings from the country's most senior traffic policeman that this would result

in an increase in deaths and injuries. Gloucestershire cancelled plans to update

cameras and has reduced or cancelled maintenance contracts.

In August 2010, the Oxford shire, UK speed cameras had been switched off

because of lack of finance due to government funding policy changes. The

cameras were switched back on in April 2011 after a new source of funding was

found for them. Following rule changes on the threshold for offering "Speed

Awareness Courses" as an alternative to a fine and license points for drivers,

and given that the compulsory fees charged for such courses go directly to the

partnerships rather than directly to central government as for the fine revenues,

the partnership will be able to fund their operations from course fees. Compared

with the same period in the previous year with the cameras still switched on, the

number of serious injuries that occurred during the same period with the

cameras switched off was exactly the same ­ at 13 ­ and the number of slight

21

injuries was 15 more at 70, resulting from 62 crashes ­ 2 more than when the

cameras were still operating. There were no fatalities during either period

3.5 Unpopularity

Use of cameras is opposed by some motorists and motoring organizations. They

have also been rejected in some places by referendum.

The first speed camera systems in the USA were in Friendswood, Texas in 1986

and La Marque, Texas in 1987. Neither program lasted more than a few months

before public pressure forced them to be dropped.

In 1991 cameras have been rejected by voters in referenda in Peoria, Arizona

voters were the first to reject cameras by a 2­1 margin. Speed cameras have

since been installed on the highways in the Phoenix area since 2007.

In 1992 cameras have been rejected by voters in referenda in Batavia, Illinois.

Anchorage, Alaska rejected cameras in a 1997 referendum

In 2002 the state of Hawaii experimented with speed limit enforcement vans but

they were withdrawn months later due to public outcry.

In 2005, the Virginia legislature declined to reauthorize its red light camera

enforcement law after a study questioned their effectiveness, only to reverse

itself in 2007 and allow cameras to return to any city with a population greater

than 10,000.

Steubenville, Ohio rejected cameras in a 2006 referendum.

In 2009, a petition was started in the town of College Station, Texas which

requested that all red light cameras be dismantled and removed from all of the

town's intersections. Enough signatures were captured to put the measure on the

22

November 2009 general election ballot. After an extensive battle between the

College Station city council and the opposing sides, both for and against red

light cameras, the voters voted to eliminate the red light cameras throughout the

entire city. By the end of November the red light cameras were taken down.

However, all citations issued are still valid and must be paid by the offenders.

On May 4, 2010 an ordinance authorizing the use of speed cameras in the town

of Sykesville, Maryland was put to a referendum, in which 321 out of 529

voters (60.4%) voted against the cameras. The turnout for this vote was greater

than the number of voters in the previous local Sykesville election for mayor

where 523 residents voted.

Arizona decided to not renew their contract with Redflex in 2011 following a

study of their statewide 76 photo enforcement cameras. Reasons given included

less than expected revenue due to improved compliance, mixed public

acceptance and mixed accident data.

3.6 Effectiveness

The town of Swindon abandoned the use of fixed cameras in 2009, questioning

their cost effectiveness with the cameras being replaced by vehicle activated

warning signs and enforcement by police using mobile speed cameras: in the

nine months following the switch­off there was a small reduction in accident

rates which had changed slightly in similar periods before and after the switch

off (Before: 1 fatal, 1 serious and 13 slight accidents. Afterwards: no fatalities, 2

serious and 12 slight accidents). The journalist George Monbiot claimed that the

results were not statistically significant highlighting earlier findings across the

23

whole of Wiltshire that there had been a 33% reduction in the number of people

killed and seriously injured generally and a 68% reduction at camera sites

during the previous 3 years.

In January 2011 Edmonton, Alberta cancelled all 100,000 "Speed On Green"

tickets issued in the previous 14 months due to concerns about camera

reliability

3.7 Avoidance/evasion

Fig 10: A GPS map showing speed camera POI information overlaid onto it

To avoid detection or prosecution drivers may:

Drive at or below the legal speed

Brake just before a camera in order to travel past its sensor below the

speed limit. This is however a cause of collisions.

Use GPS navigation devices which contain databases of known camera

locations to alert them in advance. These databases may in some cases be

updated in near real­time. The use of GPS devices to locate speed

cameras is illegal in some jurisdictions.

24

Install passive laser detectors or radar detectors that detect when the

vehicle's speed is being monitored and warn the driver. Use of these

devices may be illegal in some jurisdictions.

Install active laser jammer or radar jammer devices which actively

transmit signals that interfere with the measuring device. These devices

are illegal in many jurisdictions.

Remove, falsify, obscure or modify vehicle license plate. Tampering with

number plates is illegal in many jurisdictions.

In August 2010 a fast driving Swedish driver reportedly avoided several older

model speed cameras, but was detected by a new model, as traveling at 186 mph

(300 km/h), resulting in the world's second largest speeding fine to date behind

a man in a Swedish koenigsegg in Texas doing 242MPH!. In the past it was

possible to avoid detection by changing lanes when SPECS average speed

cameras were in use as they measured a vehicle's speed over distance in one

lane only. As of 2011 the cameras are type approved to cover multiple lanes.

3. 8 Resolving conflicts

When the edges are added to the association graph as described above, we might

possibly get a graph of the form shown in Figure. In this case, P0 can be

associated with C0 or C1, or both C0 and C1 (similarly, for P1). To be able to

use this graph for tracking we need to choose one assignment from among

these. We enforce the following constraint on the association graph “ in

every connected component of the graph only one vertex may have degree

greater than 1. A graph that meets this constraint is considered a conflict­free

graph. A connected component that does not meet this constraint is considered a

25

conflict component. For each conflict component we add edges in increasing

order of weight if and only if adding the edge does not violate the constraint

mentioned above. If adding an edge Eij will violate the constraint, we simply

ignore the edge and select the next one. The resulting graph may be sub­optimal

(in terms of weight); however, this does not have an unduly large effect on the

tracking and is good enough in most cases.

3.9 Recovery of vehicle parameters

To be able to detect and classify vehicles, the location, length, width and

velocity of the regions (which are vehicle fragments) needs to be recovered

from the image. Knowledge of camera calibration parameters is necessary in

estimating these attributes. Accurate calibration can therefore significantly

impact the computation of vehicle velocities and classification. Calibration

parameters are usually difficult to obtain from the scene as they are rarely

measured when the camera is installed. Moreover, since the cameras are

installed approximately 20­30 feet above the ground, it is usually difficult to

measure certain quantities such as pan and tilt that can help in computing the

calibration parameters. One way to compute the camera parameters is to use

known facts about the scene. For example, we know that the road, for the most

part, is restricted to a plane. We also know that the lane markings are parallel

and lengths of markings as well as distances between those markings are

precisely specified. Once the camera parameters are computed, any point on the

image can be back­projected onto the road. Therefore

26

3.10 vehicle identification

A vehicle is made up of possibly multiple regions. The vehicle identification

stage groups regions together to form vehicles. New regions that do not belong

to any vehicle are considered orphan regions. A vehicle is modeled as a

rectangular patch whose dimensions depend on the dimensions of its constituent

regions. Thresholds are set for the minimum and maximum sizes of vehicles

based on typical dimensions of vehicles. A new vehicle is created when an

orphan region of sufficient size is tracked over a sequence of a number of

frames.

3.11 Vehicle tracking

The vehicle model is based on the assumption that the scene has a flat ground.

A vehicle is modeled as a rectangular patch whose dimensions depend on its

location in the image. The dimensions are equal to the projection of the vehicle

at the corresponding location in the scene. A vehicle consists of one or more

regions, and a region might be owned by zero or more vehicles. The region

tracking stage produces a conflict­free association graph that describes the

relations between regions from the previous frame and regions from the current

frame. The vehicle tracking stage updates the location, velocity and dimensions

of each vehicle based on this association graph. The location and dimensions of

a vehicle are computed as the bounding box of all its constituent blobs. The

velocity is computed as the weighted average of the velocities of its constituent

blobs. The weight for a region Pi, Vehicle v is calculated as: is the area of

overlap between vehicle v and region Pi. The vehicle’s velocity is used to

predict its location in the next frame. A region can be in one of five possible

27

states. The vehicle tracker performs different actions depending on the state of

each region that is owned by a vehicle. The states and corresponding actions

performed by the track

28

CHAPTER FOUR

4.1 CONCLUSION AND FUTURE WORK

I have presented a model­based vehicle tracking and classification system

capable of working robustly under most circumstances. The system is general

enough to be capable of detecting, tracking and classifying vehicles while

requiring only minimal scene­specific knowledge. In addition to the vehicle

category, the system provides location and velocity information for each vehicle

as long as it is visible. Initial experimental results from highway scenes were

presented. To enable classification into a larger number of categories, I intend to

use a non­rigid model­based approach to classify vehicles. Parameterized 3D

models of idea of each category will be used. Given the camera calibration a 2D

projection of the model will be formed from this viewpoint. This projection will

be compared with the vehicles in the image to determine the class of the vehicle.

REFERENCES

A b c David Barrett (2010­08­07). "Speed camera switch­off sees fewer

accidents". London: Telegraph

Association of Chief Police Officers. 2005­03­22. Retrieved 2007­09­12.

29

a b John Lettice (2005­09­15). "Gatso 2: rollout of UK's '24x7 vehicle

movement database' begins".

Basic Concept of Operation". Sensor Dynamics. Retrieved 2008­05­31.

Bus Lane Enforcement". PIPS Technology. Retrieved 2010­04­26.

Bus lane enforcement". jai. Retrieved 2010­04­26

Freddie Whittaker. "Speed cameras will stay in Gloucestershire ­ but no more

maintenance".

HOV at Photocop.com

http://www.palisadespost.com/news/content.php?id=5646

K.D. Baker and G.D. Sullivan, Performance assessment of model­based

tracking

Liability of owner for speeding and traffic light violations".

Safe­T­Cam". Roads and Traffic Authority. May 29, 2008. Retrieved

2008­05­31.

Speed camera turn­off starts". BBC.

The Transport Manager's and Operator's Handbook 2006. Kogan Page

Publishers. p. 239. ISBN 0749444886

The Register. Retrieved 2008­07­23.David Lowe (2005).

Retrieved 2010­05­05.

U.S. DOT Red Light Camera Systems Operational Guidelines".

Vision, pp. 28­35, Palm Springs, CA, 1992.

Viper Mobile ANPR System". Sensor Dynamics. Retrieved 2008­05­31.

Wikipedia, the free encyclopedia, traffic enforcement camera.

30

31