Main Article Content
A system and method to monitor traffic in virtual mode based on surveillance cameras and the road traffic lane signal has integrated to Internet of Things (IoT).The lane traffic lights would be automated based the on frequency of moving vehicles or fixed time slots. The video analytics incorporates the machine learning for object recognition, tracking and prediction of road traffic in various scenarios. Even the moving vehicle breaks the traffic rules the automated message send to the owners mobile numbers. The system comprises of video monitoring equipment and equipment for automatic lane discipline control. The Video monitoring equipment is IoT enabled and it can facilitate traffic monitoring using remote web-based applications. The video equipment at any point in time tracks the state change of traffic signals. The real time video capturing helps to extract the vehicle identity from the vehicle number plate fixed in vehicle. The IoT enabled video camera and lane control traffic lights may be deployed at any vehicle traffic junction. The video recordings can be stored in the cloud server and applying the video analytic techniques to extract vehicle information such as type of vehicle, color and number plate information with additional attributes. Finally this system support high quality video enhancement to provide better statistical information and remote vehicle traffic control management. The GUI allows the traffic cops to operate the system automatically or manually from a remote area, thereby increasing the productivity and reducing the traffic congestion. This method is categorized into different phases such as pre-processing the video input file, applying the background subtraction algorithm to it, and then proceeding to the next operations. In this paper, we intend to design a system to classify and detect vehicles using a mainstream algorithm such as the background subtraction algorithm. For vehicle tracking we have used YOLOv3 and Simple Online and Realtime Tracking (SORT) algorithm. We have used YOLOv3 and SORT methodology for vehicle classification and extraction of information.