Since the ANPR system requires relatively advanced cameras, computing power and software, few people would guess that the technology was developed by the British Police Science Development Agency more than 40 years ago.
In the late 1970s, pilot tests were carried out on the A1 highway connecting London and Edinburgh and the Dartford tunnel. In 1981, license plate detection technology helped arrest the first car thief.
Of course, these are static systems installed on the road, not mobile systems. Due to miniaturization and technological advancement, mobile systems are generally installed in smart police cars today.
Today’s top ANPR cameras can read license plate numbers with high reliability, even on police cars moving in normal road traffic: that is, under constantly changing conditions (speed, light intensity, or changes that affect visibility Weather conditions). Regardless of the difficulties, this small detail further improves the camera's reading efficiency.
Internet connection
Cloud-based solutions—as the name suggests—require an Internet connection to work, so they may not be feasible for projects that cannot connect to the Internet.
"In the case of cloud-based ANPR, event detection is also a problem. If the bandwidth allows continuous video streaming in the monitored area (because in this case our solution uses triggers based on video analysis), it can be avoided This situation. In terms of accessible infrastructure and economic reasons, this is not yet realistic. "Another way to solve this problem is to set up an image preselection mechanism, trigger (such as an induction loop) or include A camera with a certain trigger mechanism, however, due to price differences, the latter solution may be more expensive compared to standard IP cameras that provide adequate image quality for ANPR. ”
Camera quality
is ANPR The basic rule of thumb when choosing a camera is that if the license plate can be recognized by the human eye,Then the ANPR solution can also recognize it. Having said that, because cameras on highways usually have to deal with rugged conditions, in some cases they cannot provide continuous and clear visual effects.
In all weather and lighting conditions, the camera image quality is very important. The dedicated ANPR camera system will always be better than other types of cameras because it is optimized for ANPR image quality, provides higher speed, and No image or video compression is required, which affects image quality.
The camera placement angle is also essential for capturing license plates. Too far or too close will make the image useless. Similarly, installing cameras that are too high or too low can cause problems. Finally, when the visual effect is blurred due to rain, fog, dust or other such factors, the ANPR solution will have difficulty identifying the license plate. It should be noted that these types of issues pose a challenge to any ANPR solution, whether they are located in the cloud or on the edge.
ANPR camera has not been used as a replacement for radar. However, vehicles equipped with ANPR can be relatively easily and cost-effectively upgraded to include speed detection capabilities. The cost of adding such features to modern smart vehicles is about half the cost of today's mobile vehicle radar.
In the Netherlands, ANPR is also used to identify tax evaders and confiscate their vehicles. In the UK, an extensive network of more than 10,000 static and mobile ANPR cameras collects approximately 50 million license plate readings every day, and they have detected more than 750,000 vehicles without compulsory insurance within 12 months.
TSINGSEE The license plate recognition intelligent parking system developed by TSINGSEE can effectively improve the efficiency and experience of traffic. A typical example is the EasyCVR video fusion cloud service, which has the capabilities of AI face recognition, license plate recognition, voice intercom, PTZ control, sound and light warning, surveillance video analysis and data aggregation.
The license plate recognition unit processes the image,Locate the position of the license plate, then segment the characters in the license plate for recognition, and then compose the license plate number to output. Compared with traditional card-taking and admission, license plate recognition passes without parking or opening windows, which improves the user experience of the B-end, reduces the cost of loss and loss of IC cards in the parking lot, improves the passing efficiency of the C-end car owners, and realizes the speed of the vehicle. In and out of the parking lot.
.
