Method of building a video surveillance system for searching for and tracking objects

1. A method of building a video surveillance system for object detection and tracking, the system comprising at least one camera, the method comprising the steps of:

firstly, a preparatory adjustment of each of the cameras of the video surveillance system is performed, said preparatory adjustment consisting in marking the images of the location under surveillance received from each camera of the video surveillance system by calibration elements and/or groups of calibration elements, visible in the location under surveillance, said calibration elements including segments of lines that are parallel to one another and to the horizontal plane of the floor in the space of the surveillance location, segments of lines that are perpendicular to one another and parallel to the plane of the floor in the space of the surveillance location, and segments of lines that are parallel to one another and perpendicular to the plane of the floor in the space of the surveillance location, calculation of the vertical turning angle (pitch angle) of each camera, the turning angle of each camera around its optical axis (roll angle), and the focal distance of each camera of the video surveillance system based on the calibration elements and/or groups of calibration elements, marking the polygons of visibility scopes of each camera of the video surveillance system, projecting the polygons of the visibility scopes of the cameras of the video surveillance system into a system of coordinates of the surveillance location plan, linking all cameras of the video surveillance system to the surveillance location plan, calculating the rest of the optical parameters and position parameters for each camera of the video surveillance system, calculating of the horizontal turning angle (yaw angle) of the camera is performed by rotating the polygons of visibility scope of each camera of the video surveillance system, calculation of the “x” and “y” points of the position of each camera in said three-dimensional system of coordinates of the surveillance location plan is performed by shifting the polygons of visibility scope of each camera of the video surveillance system, calculation of the height of fixture of the camera and zoom of the camera is performed by scaling the polygons of visibility scope of each camera of the video surveillance system, and plotting a coordinate match between a two-dimensional pixel system of coordinates of the video image of each camera and a three-dimensional metric system of coordinates of the surveillance location plan;

subsequently capturing the video stream from each camera of the video surveillance system with adjusted parameters of the coordinate match; subsequently frame-by-frame processing of the obtained video stream during which detecting and tracking objects of interest are carried out; subsequently plotting tracks of motion for each detected object on each camera of the video surveillance system, following which detected tracks from each camera of the video surveillance system are connected in such a way that a multicamera video surveillance is provided to the effect that the same object is transferred from one video surveillance camera to another and a single track is plotted for it in the surveillance location plan in case if such a video surveillance system comprises more than one camera; a subsequent further analysis of the nature of motion and behavior and individual parameters of such objects in all video surveillance system based on the connected tracks of the objects in the system of coordinates of the surveillance location plan.

2. The method according to claim 1 , characterized in that the method is intended for the surveillance over moving objects, for example, persons.

3. The method according to claim 1 , characterized in that, for example, one-meter long rods marked by colors of the same sequence may be used as said calibration elements.

4. The method according to claim 1 , characterized in that the polygons of visibility scopes of the cameras of the video surveillance system may be overlapped at a minimum or not overlap at all, the polygons being disposed contiguously with small gaps between polygons of visibility scopes of neighboring cameras of the video surveillance system.

5. The method according to claim 1 , characterized in that said segments of lines that are parallel to the floor are, in particular, segments of lines that lie in the plane of the floor and are of the same size in the space of the surveillance location, for example, premises; and,

physical objects-markers that lie in the plane of the floor in front of the camera in the space of the surveillance location, for example, premises, and have the same and known in advance shape, size, and color pattern for subsequently automatically detecting the position thereof and automatically determining their coordinates in the camera image.