TECHNOLOGY OF COMPUTER VISION METHODS APPLICATION FOR THE IDENTIFICATION AND DIRECTION FINDING OF ACOUSTIC TARGETS IN THE CONTROLLED AREA

Abstract

Introduction. Under the conditions of terrorist and criminal activity growth and aggravation the question about maintenance of the controlled area safety becomes one of the prominent tasks.

Currently, video camera networks are widely used to ensure the security of the controlled area. Acoustic systems for the controlled area protection have certain advantages over video surveillance.

Problem Statement. To provide situational awareness of the controlled area safety it is necessary to have accurately defined motion parameters of the acoustic targets under the condition of acoustic systems operating in passive mode, which can be achieved only with the help of indirect methods of acoustic signals estimation, received from several spreaded in space points. But such an approach creates the problem of erroneous system operation and errors in determination of a quantity of detected acoustic targets when there is more than one target.

Purpose. Development of intellectual technology of computer vision tools application for the identification and direction finding of acoustic targets in the controlled area.

Given tasks are solved with the help of such methods as processing and detection of acoustic signals; spectral and temporal analysis of acoustic signals.

Results. The tasks of stable identification, direction finding and determination of the motion parameters of acoustic targets were solved with the help of the developed technology of computer vision tools applied to sound signals, which are transformed into a stream of graphic displays. Also star clustering method was developed, which additionally validates sets of coincidence points for graphical representations of sound waves, obtained with known methods of computer vision library.