Published on : Aug 30, 2019
Researchers of South Ural State University have built up an extraordinary intelligent framework for checking flow of traffic utilizing artificial intelligence, which does not require explicit recording system and is able to function on any kind of camera. The framework, at once, analyses the information received in the real-time, not at all like existing projects where analysis causes a postponement of up to 10 to 15 minutes. Journal of Big Data has published an article on these results.
Partner lecturer at the Automotive Transport Department of the Polytechnic Institute SUSU, Vladimir Shepelev stated "We have suggested and executed a modernized framework for surveying the flow of traffic, in light of the latest advancements in the tracking and detection of the vehicles.: He has explained that in contrast to existing analogs, the new framework perceives and investigates in real-time the movement and direction of the vehicles with an optimum relative error below 10 percent. The nearest analogs can decide the speed and analyses vehicles in just a single direction and with the state of putting the cameras over the high traffic flow with a precision of 80% - 90%. Working on a neural system enables one produces up to 400 traffic parameters progressively at every crossing point.
System Provides Real-Time Data About Traffic in Every Direction
The advanced AIMS monitoring framework gathers, translates and transmits information on the power of street traffic, categorizes 10 types of vehicles, calculates speed, the present load level of every direction of the junction, and decides the further course of vehicles. Simultaneously, constant object recognition at the crossing point AIMS delivers using just one Full HD CCTV camera.
Shepelev also said, “the consequences of this investigation can be connected by city specialists to improve the general traffic limit of the cross junction.” Researchers have officially demonstrated the framework at a few cross junctions in Chelyabinsk to check that the proposed solution is adequately precise and can be utilized as a reason for other advanced level models.