ВИКОРИСТАННЯ МЕХАНІЗМІВ ШТУЧНОГО ІНТЕЛЕКТУ ДЛЯ АУТЕНТИФІКАЦІЇ ВАНТАЖНИХ ЗАЛІЗНИЧНИХ ВАГОНІВ

Анотація

Previously, intelligent technologies of transportation processes were used, which contain a number of elements of automatic data collection on transportation conditions, process modeling, compared to templates or regulations, recognition of emergencies or conditions and opportunities for their occurrence, forecasting transport systems and transportation planning. Now, for the task of recognizing freight car numbers, the mechanism of using convolutional neural networks (CNN) is best suited. Convolutional neural networks are a fairly broad class of architectures, the main idea of which is to reuse the same parts of the neural network to work with different small, local areas of inputs. The basic idea of the convolutional network is that the processing of an image area should very often take place regardless of the specific location of this area. It is the use of this network that solves a number of shortcomings and problems of many modern information reading systems, which were mentioned above.