OBJECT DETECTION PERFORMANCE INDICATOR IN VIDEO SUVEILLANCE SYSTEMS

Abstract

Context. The probability of detecting the object by the operator of the video surveillance system depends on a number of parameters (geometric dimensions of the object of observation, distance to the object of observation, parameters of the video surveillance camera, monitor parameters, etc.). Objective. The purpose of the article is to develop an indicator of the effectiveness of detecting dynamic objects when evaluating the functioning of video surveillance systems. Method. An indicator of the effectiveness of object detection when evaluating the functioning of video surveillance systems is proposed. The proposed indicator is expressed in the probability of detection of the object of interest by the i-th operator thanks to the person’s own visual apparatus or with the help of a software algorithm. This indicator differs from the existing ones by taking into account the parameters of the optical system, the parameters of the information display device (monitor), the number of video surveillance cameras, etc. The developed indicator makes it possible to estimate the probability of detection of an object by a video surveillance system operator thanks to a person's own visual apparatus or with the help of a software algorithm, depending on the distance to such an object. Results. According to the results of experimental calculations, it has been proven that the effectiveness of the use of video surveillance systems with the use of video analytics functions (using the example of the dynamic object detection algorithm). Conclusions. The conducted experimental calculations confirmed the efficiency of the proposed mathematical apparatus and allow us to recommend it for use in practice when solving problems of evaluating the effectiveness of the functioning of video surveillance systems.