VIRTUAL OBSERVER

• US 20090073265A1
• Filed: 04/13/2007
• Published: 03/19/2009
• Est. Priority Date: 04/13/2006
• Status: Active Grant

• Current Assignee: Virtual Observer Pty., Ltd. (DTI Group Limited)
• Original Assignee: Curtin University of Technology
• Inventors: Venkatesh, Svetha, Greenhill, Stewart Ellis Smith
• Granted Patent: US 9,420,234 B2
• Time in Patent Office: Days
• US Class Current: 348/148
• CPC Class Codes: G08B 13/19647 Systems specially adapted for intrusion detection in or around a vehicle
• CPC Class Codes: H04N 7/18 Closed-circuit television [CCTV] systems, i.e. systems in which the video signal is not broadcast

Abstract

This invention concerns wide-area video surveillance systems. In a first aspect the invention is a surveillance network, in another aspect the invention is a virtual observer, and in a further aspect the invention is a method for operating a virtual observer. Each segments or frame of video, from each camera of the network is associated with the following trajectory parameters; a time and a spatial position, and possibly a radius, orientation, resolution and field of view, and stored for later recall.

Claims

1. A wide area video surveillance network comprising plural cameras, wherein each segments or frame of video, from each camera of the network is stored for later recall in association with the following trajectory parameters:

a time and a spatial position.

2. A network according to claim 1, wherein the cameras are mobile cameras.

3. A network according to claim 1, wherein the cameras are static cameras at known locations.

4. A network according to claim 3, wherein the known locations are determined by GPS.

5. A network according to claim 1, wherein the plural cameras include both static and mobile cameras.

6. A network according to claim 1, wherein the trajectory parameters also include one or more of orientation, resolution and field of view.

7. A network according to claim 2, wherein the network of cameras are distributed in a fleet of vehicles.

8. A network according to claim 7, wherein each vehicle is fitted with one or more cameras having an external view looking forward, rearward or sideways.

9. A network according to claim 1, wherein a date and time stamp is recorded with each segments or frame from a local on board clock.

10. A network according to claim 1, wherein position information recorded with each segments or frame is obtained from GPS or inertial navigation devices.

11. A network according to claim 1, wherein segments or frames of a particular location are recalled from plural stored video streams and knitted together for display to a user.

12. A network according to claim 1, wherein a point in the network at which an imaginary camera is located is defined to have observation parameters corresponding to the trajectory parameters, and these are used to construct a query to recall segments or frames from storage.

13. A virtual observer in a wide area video surveillance network, comprising a computer having access to recall segments or frames of video recorded and stored, and a monitor on which it displays a view constructed from segments or frames recorded at different times, or from different cameras, or both, and associated with the trajectory parameters of time and spatial position.

14. A virtual observer according to claim 13, wherein the trajectory parameters also include one or more of orientation, resolution and field of view.

15. A virtual observer according to claim 13, wherein the view constructed is one of the following:

a view from a virtual observer given a source position, radius, orientation, resolution and field of view;

a panoramic view where the desired field of view is wider than the camera view, comprising multiple images taken at different times combined to give a wide-angle perspective view;

a “

time-lapse”

view showing how a view of a place changes over time;

a view of a particular object or landmark;

or,

a view selected on the basis of multiple spatial, temporal and geometric constraints.

16. A method of operating a virtual observer in a wide area video surveillance network, comprising the following steps:

receiving an input of the location of the virtual observer on a map of the network;

receiving an input of the query operators and observation parameters for the location;

determining trajectory parameters corresponding to the specified location, query operators and observation parameters;

recalling stored segments or frames of video associated with the determined trajectory parameters; and

,

synthesizing a view to be displayed from the recalled video segments or frames.

17. A method according to claim 16, wherein the query operators include one or more of the following:

a closest observation query that returns a segments or frame taken from the closest point to a defined location;

a view towards a place that returns segments or frames that view a defined location;

a view from a place that returns segments or frames looking outward from a defined location; and

,

a view of a large spatial region that returns segments or frames captured in the region.

18. A method according to claim 16, wherein the recalling step includes an operation to merge segments or frames of video captured using plural cameras for a particular query operator.

19. A method according to claim 16, wherein the recalling step includes an operation to merge segments or frames of video captured using one camera for plural query operators.

20. A method according to claim 16, wherein the view displayed is updated from time to time.