The History Of Detection Technology

Published Date: 02 Nov 2017

2.1 Introduction

This section, the literature review, mainly deals with the research conducted for the project in order to be able to get a better understanding regarding the problem, parking technologies and other related research.

The main focus of the literature study has been assigned to the problem, which is the cruising for parking problem. Getting a better idea regarding the problem will help build a more effective and efficient prototype. The research was also conducted on current parking technologies that are being used in current days and any new technology which is being implemented to help reduce this problem. Another section is dedicated to proposed smart parking systems, this will help get an overview of how other researchers are tackling this problem. Since the era of smart phone has increased considerably in today’s life some research has been done to investigate how mobile phone applications regarding parking are being produced in order to help solve the cruising for parking problem. Research regarding human computer interaction, reacTIVision, processing, TUIO library and image processing has also been conducted as this will be useful for the prototype which will be built after gathering the researches.

2.2 Problem Definition

Parking all over the world has always been a problem, in recent days due to the great increase in population which is leading to more cars in the streets, parking is found to be a constant struggle in everyday life. There have been various studies on parking problems mainly conducted by Donald Shoup Professor of Urban Planning (C.Shoup, 2011) together with other researchers. A few solutions on how to reduce these problems have been developed, including SF Park, Siemens Sipark SSD and Identipark whilst various other solutions have been proposed including the Image Processing Parking system and the SPS Using Ultrasonic Detector.

Parking problems deals with problems including the number of parking places available, the location and distance of the parking area, the cost charged and on street VS off street parking and other problems. One of the main problems regarding parking is the cruising for parking which is leading us to various other problems which are discussed in a later section.

On a day to day basis finding a parking space is time consuming and inefficient. Going to a particular destination and trying to find a parking space, you have to waste time going round and round the streets hoping that you would find a parking space instead of knowing how many spaces are available in the particular street beforehand.

According to (C.Shoup, 2006) the main problem is that people often go into a road and if all parking spaces are occupied, drivers must cruise to find a space which is vacated by a leaving car. Currently the only method to identify available parking spaces is by trial and error, driving slowly and making sure that you do not miss a parking space and constantly looking around and making sudden stops. Stated by (C.Shoup, 2006) (Arnott & Rowse, 2009) if each driver has a small search time for parking this will lead to create a considerable amount of traffic congestion and leads to other issues including gasoline wastage and pollution.

Various researchers have captured and analysed video tapes regarding cruising for parking and traffic flows. The researchers have also interviewed drivers who have found a parking space after cruising (C.Shoup, 2006).

A research conducted by Professor Donald Shoup at UCLA showed that between 8 and 75% of the traffic on the streets were mainly caused by drivers cruising for parking and the average time to find a parking space was between 3.5 and 14 minutes (C.Shoup, 2011), these figures vary depending on the time of day and what street the research is being conducted on.

Another research was conducted on Southbank having interviewed more than 140 drivers. From 70 on street parking drivers, 32 drivers said that they had spent more than 15 minutes cruising for parking. From the statistics it was resulted that around 90% of off street parking (parking in car parks) said that they had attempted on the streets and their average cruising time was 13 minutes (Lee, 2011).

There are various disadvantages when cruising for parking. Unfortunately when people cruise to find unoccupied parking the most obvious problem is the traffic congestion that is caused since drivers are constantly driving slowly and which is also leading to wasting time. Another problem caused is that it angers drivers which may lead to reckless driving and can cause accidents (C.Shoup, 2006). One needs to keep in mind that since this cruising for parking is increasing driving time it is leading to gasoline waste which it also effecting the environment by increasing air pollution. According to (Barata, et al., 2011) pedestrians are also affected since this also causes noise problems and retains levels of accessibility

As declared by (Arnott & Inci, 2006) during peak hours cruising for parking is increased substantially and if reducing this cruising for parking time it can result in huge travel savings and help reduces the problems that are being caused.

2.3 Detection Technology

According to (Yan, et al., 2010)an intelligent parking systems needs to have the parking space occupancy information in order to be able to help the drivers find parking. Finding the occupancy of a parking can be obtained by various types of sensor technology (Kianpisheh, et al., 2012).

From the research conducted by (Kianpisheh, et al., 2012), detecting parking space vacancy is a technology that needs detection which is able to specify if a parking space is available or occupied. The appropriate detection technology mainly depends on the objective and scope of the system (Yamanda & Mizuno, 2001). There are two main types of detection technology, vision based and sensor based.

2.3.1 Vision Based

Vision based detection normally uses cameras which is able to capture a number of parking spaces at once together with an image processing which will be able to identify the parking space availability. Vision based can be subdivided into two different approaches, one by counting the number of coming and outgoing vehicles and estimate the number of available spaces, whilst another approach is to monitor the each individual space which can also be used to guide the car to the available spaces (Kianpisheh, et al., 2012).

2.3.2 Sensor Based

On the other hand sensor based detection uses multiple sensors, normally one sensor for each parking space is used to detect each parking space occupancy individually.

The sensors are mainly divided into two categories, intrusive and non-intrusive sensors.

2.3.2.1 Intrusive Sensors

This type of sensor is mainly installed in holes on the road surface. In order to create the holes to install these sensors procedures such as tunneling under the road surface or anchoring to the roads are used. This type of sensors includes active infrared sensors, inductive loops, magnetometers, magneto-resistive sensors, pneumatic road tubes, piezoelectric cables and weigh-in-motion sensors (Liu, et al., 2012)

http://www.fhwa.dot.gov/publications/publicroads/95fall/images/p95a20.gif

Figure - Intrusive Sensor - Weight inn Motion Sensor Example (Crabtree, n.d.)

2.3.2.2 Non-Intrusive Sensors

On the other hand the non-intrusive sensors are installed by either mounting the sensor to the ceiling or ground. There is no need for tunneling or anchoring like the intrusive sensor, therefore makes them easier to install. Non-intrusive sensors include microwave radar, passive acoustic array sensors, passive infrared sensors, RFID, ultrasound and video image processing (Liu, et al., 2012).

http://www.aacomnj.com/Images/Outdoor%20Camera%20Housing%20Jersey%20City.JPG

Figure - Non-Intrusive - Video Image Processing Sensor Example (AACOM, n.d.)

According to (Yamanda & Mizuno, 2001) in order to detect the state of each space a number of methods have been utilized, including, placing ultrasonic sensors at each space which means many sensors are needed, or another method is by use of cameras which is placed at a high position and allows the guidance of a large area by a few cameras, this approach is also more appropriate for outdoor parking systems.

Different factors might be considered when choosing a sensor, including size, reliability, and adaptation to environmental changes, robustness and cost (Kianpisheh, et al., 2012).

Stated by (Kianpisheh, et al., 2012) another factor that needs to be considered is installing the sensors. For the non-intrusive sensors the installation should be a fairly easier process as there is no need for invasive procedures. On the other hand intrusive type sensors need invasive procedures to be installed which may result in traffic congestion and other problems caused by closing parking spaces until the sensors are installed.

Maintenance is also an issue when dealing with sensors, non-intrusive sensors are normally easier to maintain and replace if any damages have occurred whilst intrusive sensors might be a more demanding task which may need to close the parking space (Liu, et al., 2012).

2.4 Current Smart Parking Technologies

In this section of the literature review a few smart parking technologies that have been implemented are discussed in order to identify the different approaches and technologies that are being used in smart parking systems.

2.4.1 ParkSight by Streetline

The city of Los Angeles has started to use technology created by Streetline to help manage parking problems. Los Angeles has installed low power sensors and smart meters to be able to track the occupancy of parking spaces through certain districts.

The sensors used in the system are about the size of a coffee cup lid which is rather small. The sensors are fixed in the asphalt. This system also makes use of smart meters which are attached to the regular meters already used, these smart meters allow users to pay with their mobile phones and in addition to communicating payment information to the city (Kessler, 2011).

The sensors embedded in the asphalt are able to provide the information if a parking space is occupied of available. From the information gathered the city is able to spot the areas which are constantly full and adjust the parking price accordingly. The main aim is to adjust the price so there are always a number of spaces available (Streetline, n.d.).

The information gathered through the sensors also helps the enforcement officials by notifying them about expired parking meters or other parking violations and reduces the time they spend driving in circles to check the meters and other violations (Kessler, 2011).

Drivers are also capable of viewing the information gathered by the sensors, regarding the prices and parking availability through a free app called Parker. The application ‘Parker’ will be discussed in the next section dedicated to Mobile App Solutions.

2.4.2 Sipark SSD by Siemens

Sipark is a system provided by Siemens which is used in parking lots to be able to identify if a particular parking space is available or not. Sipark SSD uses a measuring instead of counting approach to be able to identify available parking spaces. Sipark SSD uses an ultrasound sensor to identify the availability of a parking space (Siemens, n.d.).

Every parking space has an ultrasound sensor installed on the ceiling. The ultrasound sensor is able to detect with high accuracy is the parking space is available or occupied.

The information for the ultrasound sensor is transmitted to the control center in real time. Once a driver enters the parking lot, LEDs found on the ceilings, that are able to show 6 different colours one at a time are used to help guide the users to appropriate spaces. The LED’s colours each identify something different, green means that the space is available, red means that the space is occupied and other colours are used to specify if the space is reserved for people with disabilities or any other options (Siemens, n.d.).

Sipark SSD is already monitoring more than 100,000 parking spaces in car parks found all over the world. One of the main car parks is the Munich Airport which is considered to be part of the largest parking guidance system.

Figure - Siemens Hierarchical System Structure (Parking Accessories, 2007)

2.4.3 SF Park

SFpark has been established from SFMTA which is aiming to use new technologies and policies to be able to upgrade the parking in San Francisco. By using better technologies to help users find parking this will reduce traffic in which everyone will benefit from.

According to its website (SFPark, n.d.) SFpark’s main aim is to provide safer and clearer streets. SFpark makes use of sensors which are found on each parking space, which are able to indicate if a parking space is occupied or available. Drivers are able to check the parking availability and any other important information regarding parking online, via a text message or by the use of smartphone.

SFpark operates by collecting and distributing real-time information where parking is available which therefore allows drivers to find available parking in a quicker time (SFPark, n.d.). For SFpark to be able to achieve the right level of parking availability, SFpark changes the meter and garage pricing depending on the demand (which is gathered through the use of the sensors). By adjusting the price this allows the drivers to park in underused areas which will result in less traffic congestion since the cars are spread out and availability of parking can be found everywhere (Shoup, 2011).

SFpark is testing its new parking management system at 7,000 of San Francisco’s 28,800 metered spaces and 12,250 spaces in 15 of 20 City-owned parking garages. Federal funding through the Department of Transportation’s Urban Partnership Program pays for 80 percent of the SFpark project (Shoup, 2011).

http://parkitnyc.com/wp-content/uploads/2011/10/Image_Streetline_Sensors_Graphic.png

Figure - SF Park ( Park It Guides, 2012)

2.4.4 IdentiPark

In Cavendish shopping mall which is found in Cape Town, South Africa was one of the first shopping mall’s to invest in the system ‘Identipark’ which is able to check the availability of each parking being used. The system uses ultrasonic detection sensors to detect if the parking space is available or occupied (SkiData, 2011).

Each parking space is fixed with one ultrasonic sensor on the ceiling. Ultrasonic sensors work on an algorithm known as echo-location. According to (Kianpisheh, et al., 2012) the ultrasonic sensor found on the ceiling transmits a sound, which hits a solid object, this can be the car or the ground of the parking space, then reflected back to the ultrasonic sensor. The time taken for the sent pulse and the returned echo is used to calculate the distance. If the time from the sent pulse until the echo is received is longer than the space is vacant whilst if the echo is shorter the space if occupied.

Identipark makes use of a type of LED lights which are found above each parking space to notify users regarding the current space occupation. A red light indicates that the parking space is occupied, whilst a green light indicates that the parking space is available. An orange light is used to indicates that this parking section is dedicated to parents with children and a blue light indicated that the parking section is dedicated specifically to people with disabilities (identipark, 2008).

Identipark provides drivers with an easy and convenient experience to find available parking space. The drivers are indicated to available parking spaces whilst showing the spaces that are occupied and any spaces that are reserved to parents with children and to people with disabilities.

http://www.saptakencana.com/images/arch_Identipark.jpg

Figure - IdentiPark System (identipark, 2008)

2.5 Proposed Smart Parking Technologies

This section, proposed smart parking technology gives a brief overview of new parking systems that are being proposed. These proposed ideas can help get a better understanding of technologies available to help manage parking.

2.5.1 Image Processing Parking System

A proposed solution by Ms.Sayanti Banerjee , Ms.Pallavi Choudekar and Prof.M.K MUJU have proposed a new system to provide parking information and guidance by using image processing. The proposed system includes counting the number of parking spaces that are occupied, therefore knowing how many cars are parked and identifying the available spaces.

This vision based proposed system makes use of a camera that is installed at the entry of the parking lot. The camera captures image sequences, these sequences are then evaluated using digital image processing for vehicle detection, according to the status of the parking spaces inside, real time guidance and information is provided to the driver finding a parking. When the status of the parking spaces is full no car is left to enter the parking lot. (Banerjee, et al., 2010)

2.5.2 Smart Parking System (SPS) Architecture Using Ultrasonic Detector

Another proposed solution is the Smart Parking System Architecture Using Ultrasonic Detector which is proposed by Amin Kianpisheh, Norlia Mustaffa, Pakapan Limtrairut and Pantea Keikhosrokiani. As stated in (Kianpisheh, et al., 2012) SPS is proposed to be able to assist drivers to find available parking spaces in the least time possible. This proposed system uses ultrasonic sensors to detect the parking occupancy. There have been system that track parking occupancy by ultrasonic detectors but this proposed system also checks for improper parking with these ultrasonic sensors.

Figure - SPS Architecture Using Ultrasonic Detector (Kianpisheh, et al., 2012)

SPS is able to detect the parking occupancy through ultrasonic detectors which are placed above each parking space. LEDs are able to indicate to the drivers that a particular parking may be vacant, occupied, handicapped or reserved. The ultrasonic sensors operate on echo-location. The ultrasonic sensor transmits a sound, when the sound hits an object in this case either the car of the ground it is then reflected back to the sensor. The time taken is able to indicate if the parking is free of unavailable. Another sensor on the wall is used to detect improper parking which uses the line detection system (Kianpisheh, et al., 2012).

2.6 Current Mobile Application Parking Solutions

Mobile applications are becoming more and more popular in recent years. Since mobile devices are portable and are increasing in their functionality, applications are becoming more useful and are able help the users in a number of different ways. A number of mobile applications to help solve the parking issues have been created and in this section research on these applications has been conducted.

2.6.1 Parker

Parker is an application created by Streetline. Parker application is currently available to download on smartphones, tablets and the web. This application is able to work with the system described previous, the Los Angeles parking system. Parker allows drivers to gather information which has been gathered form the on street sensors about paring availability prices and any other information regarding parking (Parker, n.d.).

Parker features include, find a place to park on-street and off-street with real-time and static information, voice guidance which lets you know where parking is nearby.

Streetline CEO Zia Yusuf says, "What we are seeking to do is build a smart parking platform, and that platform is based on data" (fankhauser, 2012).

2.6.2 Roadify

Another application that has been created to help users find parking spaces, which according to (Roadify, 2009) will eventually help reduce the time spent for cruising for parking is the Roadify application. This application is found on IPhone smartphones. Roadify does not work by sensors or any other technology but entirely on human input. Roadify allows users to enter addresses of a spot that they are about to leave or one that they happen to walk by. On the other hand, other users searching a sport in a nearby distance are now able to see the spot on the Roadify application. As specified by (Roadify, 2009) this method is based on crowdsourcing with means depending on humans to give data and when needed they get the data from other users. This application gives a good opportunity for people to share information regarding parking spots and also regarding traffic jams, navigation subway delays and buses. (Kessler, 2011)

2.6.3 Park Shark

Park Shark is another mobile phone application that is available on various different platforms. Stated by its website (ParkShark, n.d.) ParkShark’s main aim is to optimise the use of limited parking resources, reduce pollution, reduce congestion and eliminate the frustration that comes with searching for a parking space. Park Shark is a parking information and parking space sharing service.

ParkShark gives the users the ability to share parking spaces with other members of the ParkShark community in real time. Once a user specifies that there is a spot available, other members of the Parkshark community who are searching for a space nearby are able to see the spot on the application. (ParkShark, n.d.)

ParkShark uses a custom algorithm to display available spaces to those members with good ratings before displaying to other members. This encourages you to be a good ParkShark citizen and build a strong user rating. Good citizens are those that share and reserve in a reliable and consistent manner with infrequent cancellations.

2.6.4 Google Open Spot

Google Open Spot is another application created by Google. This application is based upon the user interaction. The application consists of a map with available parking spots marked with coloured dots. The dots are placed by other users when either they are leaving a parking space or when they happen to see an available one.

The colour of each spot indicates how long it has been marked by another use. Example a spot that’s 10-20 minutes old will look yellow, while a spot marked within the last five minutes will be bright red. After 20 minutes spots are removed from the map. You’re able to see parking spots within a 0.9 mile radius of your current location. (Kincaid, 2010)

2.7 Interaction Design

2.7.1 The User Experience

When dealing with interaction design, user experience is a central key point. By user experience it is mean how a product acts and how it is used by people in the real world (Rogers, et al., 2011). As stressed by Garrett (Garrett, 2010) "every product that is used by someone has a user experience: newspapers, ketchup bottles, reclining arm-chairs, cardigan sweaters."

2.7.2 Interaction Design

As defined in (Rogers, et al., 2011) Interaction Design deals with designing interactive products which are able to assist the way people communicate and interact in their everyday and working lives. Interaction Design deals with all disciplines, fields and approaches that are concerned with researching and designing computer based systems for people.

2.7.3 Process of Interactions Design

As defined in (Rogers, et al., 2011) the main process of Interaction Design deals with the following four main points.

Establishing requirements

Designing alternatives

Prototyping

Evaluating

The first step is to establish the requirements of the interactive product that is going to be built, this will help build an image of what is needed. The second step is to design alternatives which are able to show a range of different design which can be produced. The third step deals with creating a prototype based on the user requirements and the alternative designs which will give a better idea of how the interactive product will look. The fourth step is to evaluate, there are various different procedures for evaluation which can be used. The evaluation step gives the designer the opportunity to use users to get feedback and depending on the specific feedback re-iterate the prototype.

2.8 Human Computer Interaction

2.8.1 History of HCI

One important aspect that has gained a great amount of attention in recent years is the method of interaction between humans and computers, this field of study is known as Human Computer Interaction.

The ambition to build more effective weapons during World War II encouraged a great interest in the study of the interaction between humans and machines. This challenge was greatly taken up by researches of the day. (Nanni, 2004)

Since the equipment invented needed to be more effective it started to exceed the limits of human ability for safe operation, therefore the interest in applying human factors to machine became a topic of intense applied research. (Butler, et al., 1995)

Since its start from the 1990’s, "Human-Computer Interaction, sometimes called Man-Machine Interaction or Interfacing" (Karray, et al., 2008), has developed into a vast area of study (refer to the figure below) which is continuing to increase its growth during the years emphasising its the importance throughout.

Figure - HCI - Major Study Areas

2.8.2 HCI - Human Computer Interaction

Human Computer Interaction design is about helping to improve the development of interactive products by making them easy to use, effective to use and providing an enjoyable user experience. (Rogers, et al., 2011)

One aim of human computer interaction is to produce interactive products that reduce the negative aspects when using certain products whilst enhancing the positive ones and help give a pleasurable experience. When designing interactive products one main requirement is to consider who will be using the interactive products, how will they be using it and where they will be using it. (Rogers, et al., 2011)

Through the years, a number of ways of how humans are able to communicate with computers have evolved. The interface used between humans and computers to communicate is of utmost importance to help facilitate this interaction, researchers agreed that the main focus of interest should be extensive rather than simply focused on the design of an interface but should include aspects that relate to the interaction between users and computers. (Zainbooks, 2008)

One of the agreed aspects is the amount of activity between the user and the machine. (Karray, et al., 2008) identified three levels of activity:

Physical

"The physical aspect determines the mechanics of interaction between human and computer." (Karray, et al., 2008)

Cognitive

"The cognitive aspect deals with the way the users are able to understand the system and interact with it." (Karray, et al., 2008)

Affective

"The affective aspect is a more recent issue that attempts not only to make the interaction a pleasurable experience for the user, but also to affect the user in a way that would make him/her continue to use the machine." (Karray, et al., 2008)

2.8.2.1 Goals of HCI

The main goal of Human Computer Interaction is to be able to help produce an effective and efficient system. In order to achieve this HCI has two main goals, functionality and usability. According to (Karray, et al., 2008), "The actual effectiveness of a system is achieved when there is a proper balance between the functionality and usability of a system"

2.8.2.1.1 Functionality

As stated by (Karray, et al., 2008),

"Functionality of a system is defined by a set of actions or services that it provides to its users. However, the value of functionality is visible only when it becomes possible to be efficiently utilized by the user"

2.8.2.1.2 Usability

As stated by (Karray, et al., 2008),

"Usability of a system with a certain functionality is the range and degree by which the system can be used efficiently and adequately to accomplish certain goals for certain users."

Usability can be then further broken down (Zainbooks, 2008):

Effectiveness – refers to how satisfying the system is in what it is supposed to accomplish

Efficiency – how efficient is the system in supporting the users whilst carrying out their tasks.

Safety - refers to a system that is able to protect users from dangerous and undesirable situations, helps preventing users from making errors by minimising the risk of wrong and provide users with alternative means of recovery if they make errors.

Utility- refers to a system that provides the appropriate kind of functionality to allow the user to accomplish their task.

Learnability – refers to how easy and simple it is to learn how to use the system.

Memorability - refers to how easy it is to remember how to use the system one the user has learned.

2.9 Processing

Processing is an open source programming language. By open source meaning that the source code is available free of charge and easily accessible. Processing is considered to be an environment for people whose interest is to create images, animations and interactions.

Processing has started off to teach fundamental of programming but has now evolved its functionality into a tool for generating finished professional work. The Processing website (Processing, 2001) states that "Today, there are tens of thousands of students, artists, designers, researchers, and hobbyists who use Processing for learning, prototyping, and production."

Since processing is a programming language that is able to help novel technologies to be prototyped this gives a great framework for new projects that are being suggested and proposed.

According to its website (Processing, 2001) processing is used also for interactive programs, using 2D, 3D or PDF output, is able to work on various frameworks including Linux, Mac OS X and windows and offers over 100 libraries which are able to extend the software’s functionality into sound, video, computer vision and more.

Processing is able to help users by providing a community where users share code discuss processing problems and help each other build a successful product.

2.10 ReacTIVision

ReacTIVision is another open source, cross-platform computer vision framework which can be used together with processing to increase vast array of options that can be produced. ReacTIVision is used for the fast and robust tracking of fiducial markers (reacTIVision, 2005)

The main intention for reacTIVision was as a toolkit for the development of table-based tangible user Interfaces and multi-touch interactive surfaces, but its functionality has expanded and it can also be used to create prototypes of different kinds.

ReacTIVision has been developed by Martin Kaltenbrunner and Ross Bencina at the Music Technology Group at the Universitat Pompeu Fabra in Barcelona, Spain.

ReacTIVision is a standalone application, which sends TUIO messages via UDP port 3333 to any TUIO enabled client application (reacTIVision, 2005). ReacTIVision uses UDP in order to provide fast communication with local and remote client applications. (Kaltenbrunner & Bencina, 2007)

Figure - Fiducial Example (reacTIVision, 2005)

2.11 Image Processing

As stated by (Ho Kang, 2007) Image processing refers to any form of signal processing which the input is in the form of an image. The output of the image processing can result in either an image or a set of characteristics or parameters related to the image. The main goal of image processing is to be able to extract useful information from the image source.

2.11.1 Computer Vision

As stressed by (Ho Kang, 2007), computer vision deals with the science and technology of machines that are able to see. Computer vision can be seen as a scientific discipline which is mainly concerned with the theory of building artificial systems that gather their information from images. As indicated by (Ho Kang, 2007) the images used can take a number of forms, including video sequence, views from multiple cameras, or multidimensional data from a medical scanner.

"As a technological discipline, computer vision seeks to apply its theories and models to the construction of computer vision systems.

Examples of computer vision applications include:

Controlling Processes – example controlling a robot or vehicle

Detecting Events – for visual surveillance or people counting

Organising information – indexing databases of images and image sequences

Modelling objects or environment – industrial inspection, medical image analysis or topographical modelling

Interaction – as the input to device for computer-human interaction" (Ho Kang, 2007)

2.11.1.1 Feature Detection

As discussed in (Ho Kang, 2007), the concept of feature detection in computer vision and image processing refers to methods that are able to compute abstraction of image information and make decisions at every image point whether there is an image feature of a given type at that point or not.

The outcome features will be subsets of the image domain, normally in the form of isolated points, continuous curves of connected regions. Feature detection can be seen as a low-level type of image processing.

On occasions where feature detection is not feasible and there are time issues, other higher level algorithms can be used to help assist the feature detection stage, so instead of searching the whole image for features only specific parts of the image are searched for features.

Various computer vision based algorithms use feature detection as a start, therefore a number of feature detectors have been developed. The feature detection kinds vary from the computational complexity and the repeatability.

Feature detectors are mainly subdivided into the following groups:

"Edges – edges are defined as sets of points in the image which have a strong gradient magnitude. These algorithms usually place some constrain on the properties of an edge, such as shape, smoothness, and gradient value.

Corners/Interest Points – refer to point like features in an image. This algorithm uses the corners and interest points to detect the features.

Blobs/ regions of interest or interest points – Blobs provide a complementary description of image structure in terms of regions, as opposed to corners that are more point like. Blob detectors detect areas in the image which are too smooth to detect by corner detector.

Ridges – mainly used for elongated objects, a ridge descriptor computed from a grey-level image can be seen as a generalization of a medial axis." (Ho Kang, 2007)

2.12 TUIO Library

As described on the TUIO website (TUIO, 2005) TUIO is an open framework which defines a protocol and API for tangible multitouch surfaces. The TUIO protocol enables transmission of an abstract description of interactive surfaces which include touch events and tangible object states. The TUIO protocol encodes control data received from a tracker application such as reacTIVision and sends it to any client application that is able to understand the protocol such as Processing.

The main aim of TUIO was mainly for interactive surfaces, but has extended its functionality and has also been used in other related applications and prototypes.

2.13 Processing, ReacTIVision and TUIO Work Together

As stated on reacTIVisions website (reacTIVision, 2005) "reacTIVision tracks specially designed fiducial markers in a real time video stream". ReacTIVision uses the source image frame and first converts it to a black and white image with an adaptive thresholding algorithm. This black and white image is then segmented into a region adjacency graph reflecting the containment structure (Kaltenbrunner & Bencina, 2007). The graph is then searched for the unique tree structures which are embedded into the fiducial symbols. When the tree sequences is found it is then matched to a specified dictionary to retrieve the unique ID number of the fiducial. The fiducials special design is able to allow the efficient calculation of the markers center point together with its orientation. After, "OSC messages implementing the TUIO protocol encode the fiducials' presence, location, orientation and identity and transmit this data to the client applications" such as processing which has the TUIO protocol enabled. Processing can use the integers and information received and executes the processing code written by the programmer, which will then handle the according events. (Kaltenbrunner & Bencina, 2007)

By combining Processing, reacTIVision and TUIO the opportunity to create different systems has increased as these are a good framework for creating prototypes of actual systems.