Driving can be a stressful event even for the best of drivers. This in itself is felt can be made worse in busy areas like cities and large towns. In Belfast for example, the sheer number of motorists that pass through the city on a daily basis can lead to many issues arising from the congestion that ensues. Congestion on small interlinking roads cause huge problems and one of the main reasons for these congestions is illegal parking or vehicles that have overstayed the designated time limit for busy areas [1]. This may be from parking in loading bays, causing delivery men to park on the road creating an obstacle for other motorists or parking on yellow lines, footpaths, cycle lanes etc. The end result is that busy roads become increasingly difficult to navigate which may result in congestion. This needs to change. Life could be made easier if a system was developed for places like Belfast to monitor parking. However a system like this shouldn’t be seen as a punishment but more to be seen as a learning measure to benefit all motorists.
This topic warrants investigation as a similar scheme involving use of bus lanes was introduced within Belfast. Cameras were installed to monitor the illegal use of bus lanes. A system like this was needed as in Belfast motorists were found to be using the bus lane when it was signed clearly that at certain times buses were the only vehicles to be using those lanes. Motorists that were caught illegally using these lanes by the ANPR system were caught and issued with tickets/fines. This in turn dramatically reduced the number of motorist using these lanes as a fast track route and in turn the number of fines issued. A result like this with parking is what this system is looking to achieve as it will benefit not only the government as it will reduce the amount of people roaming the streets to catch motorists but also the motorists as the system will be enough of a deterrent itself as.
This system is aiming to eliminate ‘over staying’ in parking spaces and illegal parking by developing an Automatic Number Plate Recognition (ANPR) automated ticketing system that will aim to identify illegally parking cars in either non-parking spaces or in a designated parking space to which the time limit has been exceeded. This will work by identifying the spaces, the car make, type, colour and then number plate associated with that car linking it to the registered owner. Once the car and owner have been identified the parking ticket/fee will be sent to the registered address when normal procedures will take place. This system will aim to reduce the amount of illegal parked within some of the busiest areas of cities thus, this system will reduce congestion and promote free flowing traffic. An automated system will act as a better deterrent in these cases as parking tickets are automatically issued in every instance.
1.1 Aims and objectives
This project aims to:
• Identify legal and illegal car spaces on a road.
• Identify vehicle type colour and number plate, linking the vehicle to the registered owner.
• Recognise when a vehicle has in fact stopped in a space through use of a set time the vehicle is stopped, for example if a vehicle is found to be stationary for over a minute they are considered parked.
• Know which spaces have limits for parking, which or disabled, loading bays or not spaces at all in the street in view.
• Locate and track vehicles once they enter the view of the camera.
• Reduce congestion on the roads by reducing the illegal parking on roads
• Once a car is found to be parked illegally a ticket is automatically issued to the vehicles registered owner’s address along with evidence to why it was issued.
The main objectives to this project are to create an efficient system that locates and track cars that enter the frame. A link will be made between the number plate, car make and colour to its registered owner, making full details available to the authorities in the event that a ticket is to be issued. This will over time, reduce the amount of tickets issued for illegal parking. This will be achieved by eliminating the chance it factor which is associated with parking when it’s parking wardens, as the ANPR parking ticket system will always catch the person who is in a parking space over the limit or if they shouldn’t be there. This will result in less people doing it, resulting in less tickets being issued, saving the driver money but also the government as this will in turn result in less or no traffic wardens needed to be walking certain streets and allow these resources to be allocated elsewhere.
1.2 Literature review
Within this chapter each element of the project will be researched and investigated. For this project, this is existing ANPR systems, companies that supply ANPR service for parking; how these companies do this and what elements would be beneficial to this project. Tracking algorithms are also important to this project, the reason for this is that the system will need to track vehicles in order to determine location of the car, if the car has stopped and if the car has stopped. Database management systems will be explored as this system will require the storage of data and evidence, a database will store information of registered owners, evidence and information of offences. Also the possibility of a GUI being incorporated into the system to allow ease of use for the user.
1.2.1 Background
In 2016, 85,000 on-street parking tickets were issued across Northern Ireland. It is estimated by the Northern Ireland Statistics and Research Agency that in 2016 the population stood at 1.862 million of which only were able to drive. That being said not all drivers own cars [2]. Nevertheless, it is felt that this demonstrates that on-street illegal parking is a huge problem across Northern Ireland and that the current system in place isn’t enough of a deterrent to pervert drivers from parking illegally. This is due to the same people willing to take a chance that they won’t get caught. Having obtained the figures in relation to on-street tickets issued, there is a need for a new protocol to be put in place. It is proposed that an introduction of ANPR system will assist in reducing illegal parking and as a result congestion. This system will identify vehicles by license plate number; vehicle type and colour. The system will use ANPR and a tracking algorithm to identify and follow cars that enter the frame. Once the vehicle has been registered as stopped/parked in an illegal space the system will connect the vehicle to its registered owner sending them an appropriate ticket.
This will eliminate the possibility of not getting caught therefore acting as a greater deterrent than the current system that is in place.
1.2.2 Parking Enforcement
ANPR Bus lane cameras Belfast
Statistics released by the Department for Infrastructure show that fines for driving in bus lanes with 32,254 tickets handed out in 2015 for ‘moving traffic offences’, all being issued in the Greater Belfast area. The fines only covered the period from June 22 to December 31, 2015. The number of tickets issued fell significantly in 2016, with 27,911 tickets issued over the entire year. From January to September 2017 the total number issued 15,356 this shows that the number of PCN’s issued is continuing to decrease, this is a direct result of the ANPR cameras for bus lanes that were installed in Belfast to combat the illegal usage of these lanes by motorist. The company trusted with the ANPR bus lane operations is MAV ANPR Solutions. They offer a wide range of ANPR solutions from Car park management to Smart city. MAV utilize ANPR to detect any motorists using bus lanes illegally as shown above by the statistics of PCN’s issued in Belfast these lanes are often misused by drivers resulting in the possible bigger problem of congestion, increases the likelihood of accidents and also puts the safety of other drivers and pedestrians alike at risk. This ANPR system has been very successful since being incorporated into Belfast. it identifies vehicles using bus lanes during illegal periods thus issuing the vehicle owner with a ticket accordingly. This is an aim of the system proposed only for illegal parking and illegal use of spaces. The ANPR Parking ticketing system will incorporate features from this system, one feature is the timing. Like with bus lanes, parking spaces have times during the day where you can or can’t park in a space, spaces also on different days or times during the day have different time limits which will have to be featured in this proposed system or else the system will be issuing false tickets [6].
Serco
Serco Group plc offers a wide range of services from Health to Citizens Services. One of which is parking enforcement. Serco parking enforcement cover car park management, controlled parking zones, and moving traffic offences across the road network. Serco’s parking enforcement is obtained by using ANPR enabled teams, through the use of street segmentation and analytics. This allows Serco’s enforcement resources to be precise and focused letting them be on hand where and when needed.
In London, Serco’s parking enforcement methods have had a great success, even with an increase of 0.9 percent across London which is expected when CCTV and ANPR parking enforcement systems are put in place. Initially in Hounslow, where the system has been established there was a 10.3% decrease [].
The use of Serco’s ANPR parking enforcement system has demonstrated a reduction in illegal and dangerous parking thus reducing traffic offences. This system reduces congestion on roads in turn improving traffic flow for visitors, local residents and businesses. This is the main aim for Serco’s system; this is also one of the main aims for the ANPR Parking ticketing system. Serco’s parking enforcement systems incorporate a lot of the aspects that the ANPR parking ticketing system will incorporate. Serco segregate the streets by parking and non-parking accordingly similar to the way this project will approached. [3]
Fig 2. PCNs issued in London since Serco ANPR introduced
Redflex
Redflex Holdings is a company that offers vehicle monitoring and enforcement services for government, police, and traffic departments globally. They produce a wide range of digital photo enforcement products and systems, utilizing advanced sensor and image capture technologies.
Redflex use ANPR camera technology with ‘REDFLEXexpress™’ software to capture the number plate details of vehicles of interest. As a vehicle passes the camera, Optical Character Recognition (OCR) technology matches the vehicle against the linked database for its owner. This Redflex system can be deployed as a standalone ANPR unit or can be interfaced with other REDFLEX speed and/or red light systems. The service which is of main relevance to this project is their ticketless parking where they have developed an ANPR system using cameras which have the ability to capture all cars parking and leaving a designated parking area, whether it is on or off street. The system provided by Reflex identifies the License plate of the vehicle on entering the parking zone and leaving then issues a ticket accordingly. [4]
This system has similar aspects that will be incorporated into the proposed ANPR Parking Ticketing system in terms of the ANPR, timing and tracking aspects of Redflex ANPR systems however this system is for parking in a designated space for an amount of time and the proposed ANPR Parking Ticketing system is for illegal on-street parking.
ANPR Parking Control
ANPR Parking Control (ANPRPC) is a company that offer an array of parking revenue systems around the UK. They provide ANPR system for cars parks mainly aimed at supermarkets, retail parks & large retailers. This systems main focus is to reduce or stop noncustomers taking up their valuable customer parking spaces which causes a loss in revenue for the outlet or outlets.
It has been estimated that up to 15% of parking revenue for business’ and retail parks is lost on an annual basis this is due to ticket fraud.
ANPRPC works through the use of ANPR cameras utilizing tracking software which track entry and exiting of a vehicle which are linked to a wireless GPRS, once the vehicle is found to have left the parking structure, the ANPRPC system automatically issues a Parking Charge Notice, delivered by Royal Mail to the registered keeper of the vehicle as held on the DVLA database to all vehicles that do not abide by the parking restrictions as advertised on the Warning Signs in the parking areas i.e. maximum stay, etc [5]. The system and service provided by ANPRPC is similar to the proposed ANPR street parking ticketing system discussed previously, the only difference is that ANPRPC only offer service to mainly car parks and companies that would have car parks where street parking is the aim of the ANPR street parking ticketing system. The system offered by ANPR has a lot of similarities to the system proposed.
1.1.3 Automatic Number Plate Recognition (ANPR) Systems
So far throughout this report the discussion has revolved around companies and examples of times that ANPR software and camera have been utilized within companies to tackle parking issues for both on-street and off-street parking. This section of the report will look at ANPR open source systems and open source code.
OpenALPR
OpenALPR is an open source Automatic License Plate Recognition library written in C++ with bindings in C#, Java, Node.js, Go, and Python. The system analyses images or video streams identifying vehicle number plates, the output of which is the text representation of any license plate characters. Once the image or video has been analysed the system will give a breakdown of the car over these categories: License plate number; vehicle make; vehicle model; vehicle type; vehicle colour and processing time. The system also gives the level of accuracy of each of these categories and the seconds it takes to process which is usually less than one second. [7]
This system is a good starter to base the project off as it has all the features that is need for the proposed system in terms of ANPR. It is an accurate system that can be used for still images or videos which is ideal to be used as part of a system that uses live ANPR cameras to identify illegal parking. This system also has a fast processing time usually under one second which is required for the proposed system as the ANPR Parking Ticket System will need to process in real-time.
1.2.3 Vehicle Tracking Algorithms
Vehicle Detection, Tracking and Counting
Fig 3. Background Subtraction
There are various ways to perform vehicle tracking. One is to use a Background Subtraction (BS) algorithm. BS algorithms also known as foreground detection is a commonly used method for detecting moving objects within a video captured by a camera. BS works by identifying the foreground objects of interest in a frame; these are usually cars, people, text etc this foreground image is achieved by subtracting the current frame and a background frame, the current frame is the original frame and the background frame is the frame without the object or objects of interest.
The Vehicle Detection, Tracking and Counting is an open source vehicle tracking written in Python and C++. This software uses BS to track vehicles over a certain distance within a video. The video is segregated once a vehicle or vehicles enter that section they are detected, tracked and counted.
This system would be a good base for tracking vehicles for the proposed ANPR Parking Ticketing system. It has features that are needed for the system in terms of tracking and detection. Vehicle counting isn’t a feature that has been considered for the system however could be beneficial. It is an accurate system that tracks objects in a video feed making it great starting point for the proposed system [8].
Vehicle detection, tracking and counting
This is an open source vehicle detection and tracking algorithm. This algorithm has been written in Python and C++. It uses an API and has libraries from OpenCV3.x. This vehicle tracking, detection and counting software uses a background subtraction algorithm to allow vehicle detection to occur. The below images show both the input video and the video once background detection has taken place
In order to track the vehicle a tracking algorithm is used. The specific one used in this software is a blob tracker algorithm using the cvBlob library which the foreground mask is sent to, this library provides methods to get the centroid, the track and the ID of the moving vehicle. Below is the tracking feature in the system and the vehicle counting feature.
1.2.4 Databases
Relational Database:
A relational database is a collection of data that is organised in a table format, which is then structured and relationships formed between the tables, see Fig 4 for an example of a relational database set up. The data can be accessed and edited in a various number of ways without the need of reorganising the tables themselves [9]. There are many advantages to using a database management system like this; such as: a relational database is easy to use and would require little to no training to use; database queries can search any column for matching entities; it allows multiple users to access the database at the same time. This is due to the built-in locking and transactions management functionality. This feature allows users to access data as it is being changed, preventing collisions between users as they edit or update data. This then keeps users from accessing partially updated records. Relational databases allow the database administrator to restrict the access certain users have to the database but also grant privileges to individual users based on the types of database tasks they need to perform and the relational database is easily maintained due to its built-in maintenance utilities and built-in automation meaning many of these functions can be automatically performed once set [10]. However, the relational database is not the fastest database type but can be improved by increasing processor speeds and decreasing memory and storage costs allow systems administrators to build incredibly fast systems that can overcome any database performance shortcomings.
Hierarchical Database
A hierarchical database is a data model where the data is organized into a hierarchy/ tree-like format, see Fig 5 for an example of a hierarchical database structure. The data in a hierarchical database is known as records which are stored then connected to one another by links. A record is fields, every field can only contain one value. The entity type of a record defines which fields the record contains.
In the hierarchical database model, records correspond to a row within the relational database model, an entity corresponds to a table. In the hierarchical database each node is referred to as a parent node or child node [11]. The hierarchical database model mandates that each child record has only one parent, whereas each parent record can have one or more child records. In order to retrieve data from a hierarchical database the whole tree needs to be traversed starting from the root node. There are many advantages to using the hierarchical database: it allows the user to easily add and delete information; Data at the top of the Hierarchy database is fast and easily accessed; the hierarchical database works well with anything that has a one to many relationship, however there are also disadvantages; the hierarchical database requires all data to be repetitively stored in many different entities; it can be a slow database model when searching for information on lower entities; linear data storage is no longer used; searching for specific data this requires the database to search through the whole model top to bottom until the data that has been searched has been located. This will then in turn make queries extremely slow. Within the hierarchical database system the user can only model one to many relationships as many to many relationships are not supported.
2018-4-25-1524669168