by:Shenzhen TGW Technology Co.,Ltd
Automatic number Plate Recognition (ALPR) is a technology that can detect the Vehicle on the monitored road and automatically extract Vehicle License Plate information (English letters, Arabic Numbers and Plate color) for processing. License plate recognition is one of the most important parts of modern intelligent transportation system. Based on digital image processing, pattern recognition, computer vision and other technologies, it analyzes the vehicle images or video sequence taken by the Anpr camera to obtain the unique license plate number of each vehicle, thus completing the recognition process. Parking fee management, traffic flow control index measurement, vehicle positioning, car theft, automatic superhighway overspeed supervision, red light electronic police, highway toll station and other functions can be realized Because of a serious of follow-up processing . It is important for maintain traffic safety, fighting crime,prevent traffic jam and realize traffic automation management.
In fact, we can pinpoint the culprit of many traffic violations or other crimes by tracing to the license plate. Many intersections are equipped with a red light violation detection system, and the most important part of this system is intelligent number plate recognition system modules, the system record the number plate of vehicle , and then analysis which are suspicious vehicles according to Database comparison, determine illegal information, thus reducing the illegal behavior, control the status of the traffic on the road to ensure safety.Countries around the world are actively using ANPR to combat crime.
On April 24， 2015,it was announced that Victoria Police to consider implementing $86 million live video ANPR system for patrol cars
in November 2014, it was announced that The Royal Malaysian Police (PDRM) was to employ an Automated Number Plate Recognition
(ANPR) system to identify vehicles with foreign registration plates with unsettled summonses
On December 20, 2018， the local authority of Redbridge Announced that £1.5m ANPR cameras installed to fight crime.
Automatic number plate recognition software is proving invaluable to the Huntly police force.Within four hours of using the ANPR
equipment last week, the team of eight stopped over 25 vehicles, including a car which had been stolen by knifepoint in Northland.
Nowadays, more and more parking lots have improved access to ANPR to improve management efficiency. The traditional parking system cannot meet people's needs because of the high labor cost and low management efficiency. The parking lot managers are planing to update the traditional parking systems into smart parking systems
The traditional parking system can be modified by the automatic number plate recognition (anpr) technology. The ANPR camera captures the vehicle information and the image of the person clearly, so that the owner can enter and exit the parking lot without stopping. It can also monitor the running status of all online devices, realize equipment fault pre-discovery, self-repair, query various report functions in real time, and provide customers with on-site information in the form of multiple pictures, with real-time query charge amount, the system breaks through the region. Sexual restrictions have enabled remote management across the country. The parking lot charging system is charged automatically and supports the current popular mobile payment or login to a special parking app for payment. The whole process is fast and convenient.
At present, the traditional parking lot management system is prone to cause congestion during the peak hours of commuting. The traffic efficiency is extremely low because of the simple function of the system. Traditional parking lot does not have the functions of helping the owner to find a parking space, finding a car and paying a fee quickly. Manual charging also leads to capital loopholes, makes difficulty in collecting money for change, etc. it is necessary to install a separate database and front-end display for the parking lot during installation. The ANPR system adopts a fully automatic computer management mode. From the exit of the parking lot to the gate and then to the management center, all of them are maneged by the automatic control computer to make it convenient to manage. ANPR decrease the time we spend on vehicle entering the parking lot and leaving the parking lot. And it can easily realize unmanned operation at the entrance and exit, and automatically charge, which can greatly save costs. Therefore, the intelligent parking system that can meet the needs of the management and the owners has very high potential, and the future development prospect is very broad.
The basic hardware configuration of a license plate recognition system is composed of A ANPR camera, master controller, acquisition card and lighting device. The software is composed of an image analysis and processing software with license plate recognition function and a background management software to meet specific application requirements.
License plate recognition system has two sorts of products, one is a combination of hardware and software, or the realization of hardware recognition function module, forming a full hardware license plate recognition, such as DSP. Another sort is open software and hardware system, that is, hardware adopts standard industrial products, software as embedded software. Both the two product formats have advantages as well as disadvantages. The advantage of the open system is that the hardware adopts standard industrial products, and the operation and maintenance are easy to grasp. The purchase of spare parts can be obtained from any manufacturer, and there is no need to worry about the permanent failure of products or procurement difficulties due to the failure or shortage of supply of one manufacturer. The integrated hardware and software products make it easier for users to operate and control the products. For later maintenance debugging is also easier to grasp.
Automatic number plate recognition is a pattern recognition technology that USES dynamic video or static images of vehicles to automatically recognize license number and license color. Its hardware base generally includes trigger equipment (monitoring whether the vehicle enters the field of vision), camera equipment, lighting equipment, image acquisition equipment, processor (such as a computer) that recognizes the license plate number, etc. Its software core includes license plate positioning algorithm, license plate character segmentation algorithm and optical character recognition algorithm. Some ANPR systems also have the ability to determine whether there is a car by using the video vehicle detection techno. A complete license plate recognition working process(as shown in figure ) should be include of vehicle detection, image acquisition, license plate recognition and other parts . When the vehicle detection part detects the arrival of a vehicle, the image acquisition unit is triggered to collect the current video image. The license plate recognition unit processes the image, locates the license plate position, then separates the characters in the license plate for recognition, and then output the license plate number.
There are three indexes to evaluate a number plate recognition system: recognition rate, recognition speed and background management system. Of course, the premise is that the system should be able to operate reliably.
Whether a ANPR system is practical or not, the most important index is the recognition rate. The international traffic technology has made the special identification rate index elaboration, the request is 24 hours all-weather brand accurate identification rate above 95%.
In order to test the recognition rate of a number plate recognition system, it is necessary to install the system in a practical application environment, operate 24 hours a day or more, collect license plates of at least 1000 natural traffic flows for identification, and store license plate images and recognition results for retrieval and viewing. Then, it is necessary to get the image of the actual passing vehicle and the correct result of manual identification. Then the following recognition rate can be calculated:
1. The identification rate of natural traffic flow = the total number of correct identification of the full plate/the total number of actual passing vehicles
2. The percentage of identifiable license plates = the total number of license plates correctly read by human/the total number of vehicles actually passed
3. The correct recognition rate of all plates = the total number of license plates correctly recognized by all plates/the total number of license plates manually read. These three indicators determine the recognition rate of license plate recognition system.
Recognition speed determines whether a license plate recognition system can meet the requirements of real - time application. A system with a high recognition rate, if it takes a few seconds or even minutes to recognize the result, will not be practical because it cannot meet the real-time requirements in practical applications. For example, speed is a powerful way to reduce travel time and avoid lane congestion in an application that USES license plate recognition to reduce travel time.
The speed of identification proposed by international traffic technology is less than 1 second, the faster the better.
The background management system of a license plate recognition system determines whether the license plate recognition system is useful. It is important to keep in mind that it is impossible to achieve a 100% recognition rate because the license plate is defaced, obscured, obscured, or the weather may be bad (snow, hail, fog, etc.). The functions of the backend management system should include:
1. Reliable storage of identification results and vehicle image data, which can protect image data from loss when the multi-function system operation makes the network error-prone, and facilitate manual investigation afterwards;
2, effective automatic alignment and query technology, the recognized license plate number with the database of thousands of license plate number automatic alignment and alarm prompt, if the license plate number is not read correctly when the use of fuzzy query technology to obtain the relative "best" comparison results;
3, a good license plate recognition system for network operation, but also need to provide real-time communication, network security, remote maintenance, dynamic data exchange, automatic database update, hardware parameter Settings, system fault diagnosis.
1.Attaches great importance to stability
With the improvement of product algorithm and industry technology, integration is not an important topic for a long time, but the most important point at the present stage is "stability". Stability is defined as: within a given speed range, the accuracy must be achieved without excessive error due to external environmental influences.
For example, a license plate system that is more than 90 percent accurate during the day, down to 80 percent in the evening and 70 percent at night is more difficult to integrate than a license plate recognition system that is on average 70 percent accurate throughout the day. Because the user might think, now that the recognition rate of 90% during the day, the all-weather only reasonable accuracy is 90%, this specification does not include strange environment interference (torrential rain, hail, fog, sections, etc.), and set up the environment limit (, big wind shake height limit restrictions, it is not easy to artificial damage, etc.).
Almost every company claims to have high recognition rate, but in order to avoid later because both sides cognitive difference of product, and to each other, the responsibility of dysfunctional users in purchasing the license plate recognition system, as well as for product testing, and test had better more than two weeks time, compare results can determine whether "exaggerated". Because of the changeable environment, two weeks should be able to grasp about 80% of the conditions that may affect the identification rate of the field. If you only measure it for a day or even a few hours, you cannot understand it.
In addition, since license plate recognition is a "system", the hardware and software architecture will certainly affect the "present results". As to what kind of software and hardware are suitable for what kind of environment, it must be different from environment to environment, because different application environments may not have the same requirements on identification rate, and this must be accumulated by experience.
Although there are a variety of license plate recognition system on the market, the use of the product and structure, can save a lot of wasted money and time, but more importantly, the project and system integrators need to cooperate with and understand, rather than blindly focus on a brand is better, cheaper, nothing is cheaper.
In addition, whether the license plate recognition system can make the most use of the software technology, the camera and the field construction ability, also have a great deal to do. Users may request the manufacturer to the scene after the exploration, the paper puts forward establishing planning, the assessment should set up the first place, using an Angle, whether you need to set up auxiliary light source, etc., offer again, by these actions, in addition to assess in advance ability and users themselves can also achieve product learning and education training, management, in the future will be more clearly know the restrictions on the use of the products and the relevant measures.
More info in wikipedia