Development of an Application for Recognizing Automobile Vehicles

In modern cities, a busy traffic is an acute problem. Therefore, road monitoring should become an essential feature of smart cities. The possible approach to car traffic control is the usage of neural networks. In this article, we consider the usability of this approach for the detection and classification of city traffic. The developed application is based on the convolutional neural network of the YOLOv3 system. With the testing data set of 750 images, the trained neural network showed an accuracy of 76%.