A new method for automation of the personnel authentication process using a video stream

On the basis of research carried out on automated control systems for industrial enterprises (ICS), the work proposes a mathematical model for the authentication of personnel using the technology of generation and processing of a heterogeneous stream containing photographic identification signs of personnel from video surveillance sources in the traffic of video data. A distinctive feature of the model are navigation operators and outputs to signal processing according to aggregate stream video management, allocation of frame-inserts bearing photographic identification signs, segmentation of space images required for authentication of staff in integrated access control system, as well as the use of technologies to restore the video stream. A technique for identifying frames-inserts in heterogeneous video stream has been developed, which is based on the estimation of total differential extreme brightness of frame-folds and an analysis of the histograms of image frames according to the correlation coefficients and polynomials of Chebyshevskii approximation. This technique uses dynamic threshold for authentication of frames and making a decision about their class on the basis of a majority method of data processing. The paper presents the results of simulation of the process of personnel authentication on the basis of the proposed model and methodologies, allowing one to estimate efficiency of visual personnel authentication, meeting the requirements of the reliability of decision making and limiting the transmission capacity of the data network.