Researchers in Iran Design System to Detect Vehicles License Plates Using AI
Mohammad Mahdi Ebrahimi, the managing director of a knowledge-based company in Iran said in an interview with ANA that their firm is operating in the field of artificial intelligence and producing image processing software. "Cars traffic control software, weighing scales software, using modern programming techniques for managing databases are important products being worked on and produced in this company," he said.
Ebrahimi also said that the parking lot management software is considered as one of the most comprehensive software in the country, adding, “This software not only is easy to use, but it ensures the security of vehicles well with the capability to automatically compare cars when entering or leaving the parking lot."
“In this system, identifying the vehicle is done by automatic license plate detection, barcode card, RFID card, long range antenna or a combination of them,” he added.
The managing director of the knowledge-based company further said that the detecting license plate number is one of the most suitable information data-x-items to identify the vehicles, adding, “This automatic system can monitor and manage entry and exit on its own.”
According to him, the system takes an image of a passing car and then detects the license plate number using advanced image processing algorithms.
“This software has the capability to determine types of traffic permits, connect additional cameras, control traffic lights, and operate based on the network (internet),” Ebrahimi explained more.
Stating that the software takes pictures of the vehicles with the cameras that are installed at the location, he continued,” The license plates are translated into text using artificial intelligence algorithm methods of imaging and reading license plates. Registered license plates are also used in the system.”
“License plates registered in the system are matched with the list of authorized and unauthorized license plates. This system is capable of displaying all registered traffic separately on a specific day,” the researcher further said.
“Using the roadblock control circuit and connecting it to the traffic control software, intelligent control of roadblocks, electric doors, etc. can be mechanized,” Ebrahimi also said.
“The license plate reading core has been implemented by this software with 97% accuracy. It also has the capability to continue its services offline,” Ebrahimi went on to say.
The managing director of the knowledge-based company further said that almost 300,000 characters have been introduced to the system for training which makes it be capable of detecting license plates with high accuracy.
Ebrahimi further said that their locally produced system is sold for much lower price than imported foreign-made equivalents.
4155/i