Optical character recognition (OCR), which enables each group of pixels in an image to be scanned and then determines whether the group constitutes a letter or a numerical figure and if so converts the pixels to the appropriate ASCII code for that letter or figure, is the foundation stone of automatic number plate recognition (ANPR).
ANPR is poised to become a pervasive technology primarily because OCR also conveniently provides the ability to both store and sort data. This means that images captured by ANPR cameras can be scanned and the vehicle registration number converted to ASCII code and stored in a list. Depending on requirement, an ANPR system will scan up to 15 images and select the best for conversion and storage in the list.
Pitfalls
Some of the pitfalls that new ANPR technology strives to overcome include blurring of the number plate characters as a result of poor focus or vehicle movement and acute presentation because of poor camera angles and low contrast. The numbers can be obscured because of vehicle headlights or reflections from the sun and many other factors may also influence the clarity of the images.
Conventional CCTV cameras, therefore, cannot always be successfully implemented in the role of ANPR because there are so many variations that affect lighting conditions, not the least of which is the retro-reflective nature of modern number plates that results in light being reflected directly back to the light source.
Because of this, special ANPR cameras have evolved that are sensitive to an infrared illumination source which eliminates other light reflections or refractions in both daylight and under other illumination at night. They are also capable of shutter speeds of 1/100 000th of a second to capture the number plate detail of fast moving vehicles.
Other considerations
While OCR and ANPR cameras constitute the operational guts of any ANPR access control system, a lot of other factors have to be taken into account including the size of the number plates in general use, the different shapes (both rectangular and square shaped number plates are in use in South Africa), camera lens selection and the distance at which vehicles will be viewed when the images are captured.
Bearing in mind that for OCR to function correctly, the character detail of the number plate as captured by the ANPR camera has to be of a certain size and the vertical height is important. For an application such as control of parking in the basement of a large corporate headquarters, the camera placement, lens angle, distance from the entering vehicle and other parameters would be calculated prior to system installation in order to achieve the best result.
In a typical parking control application, shutter speeds of 1/100 000th of a second are not necessary because the speed of approach to the barrier is generally low and therefore 1/250th of a second would in all likelihood suffice.
Number plates from the various provinces also display different symbols in addition to character and figures. Typically, Gauteng number plates include the provincial coat of arms. Other provinces have similar embellishments and the ANPR system can be configured to be able to read these as an integral part of correct identification.
Depending upon the degree of access control/security required, a colour camera can be added below or adjacent to the ANPR camera system to clearly identify the colour and type of vehicle as well and this image can also be stored in the database and linked to the number plate file. This provides a useful means of quickly checking against false number plates.
Data storage and access
Thought needs to be applied to the database from a number of viewpoints. The design of the screen is important, as it should show the colour photograph of the vehicle, the ANPR image of the number plate and the list of vehicles that have entered/departed the premises.
The list will display the ASCII information - number plate, time of entry, date of entry, the category or status (managing director, contractor etc) of the driver and the point of entry. For example, if the location using ANPR is a major retail goods distribution hub, then the status of the driver would display goods delivery/goods collection and identify the transport contractor/employer and the point of entry - main gate, east gate etc.
Functionality
Systems can be configured to give a range of functionality to suit different applications and needs. These include:
* The ability to add notes to the image files, determination and display of the length of time a vehicle has been on site (this could be linked to an alarm flag if the usual turnaround time is exceeded).
* Employee names linked to number plates and access privileges that will automatically allow access or deny access to certain areas, depending on the privileges set for that person.
* The provision of a full list of all the vehicles and their registration numbers on the site at any given time, the software separating out from the list those vehicles that have left the site at the time the list is processed.
* Drivers are often associated with one vehicle so biometrics (face or fingerprint) or a personal tag such as a ticket or ID-tag could be applied on exiting to improve security against theft. Company vehicles driven by more than one person can have multiple biometrics, or tags, identifying a number of approved drivers.
* In pay parking facilities at shopping centres, airports and parking garages in commercial centres, the number plate can be used to automatically enter regular pre-paid parkers who may pay a monthly fee. For casual parkers, the system can also be used with pay-on-foot systems or can calculate parking fees at manned exits by comparing the exit and entry times.
Future developments
Use of number plate recognition technology has expanded in recent years and is now poised to grow exponentially. Some of the triggers for the technology include traffic congestion and control, rising vehicle theft and the need for secure vehicle parking at businesses, office parks and public facilities.
There is also potential for the linking of ANPR systems with stolen vehicle details from the police so that when such a listed stolen vehicle enters an ANPR controlled parking area it can be flagged as such and the relevant authorities notified.
Future potential development includes the control of vehicle flow and access to central business districts and highways, bearing in mind ever increasing volumes of vehicles which are forcing authorities in many parts of the world to seek ways of controlling and/or limiting access for private vehicles at certain times in an attempt to reduce traffic congestion while promoting use of bulk public transport.
For details information contact Dr Bennie Coetzer, Protoclea Advanced Image Engineering, +27 (0) 11 465 4312, benniec@protoclea.com
© Technews Publishing (Pty) Ltd. | All Rights Reserved.
printer friendly version