System and method for matching image information to object model information

• US 6,185,314 B1
• Filed: 02/06/1998
• Issued: 02/06/2001
• Est. Priority Date: 06/19/1997
• Status: Expired due to Term

• Current Assignee: FLIR Commercial Systems Incorporated (Teledyne Technologies Incorporated)
• Original Assignee: NCR Corporation (NCR Voyix Corporation)
• Inventors: Moed, Michael C., Crabtree, Ralph N., Khosravi, Mehdi
• Primary Examiner(s): Tran, Phuoc
• Assistant Examiner(s): Mariam, Daniel G.
• Application Number:
  US09/020,113
• Time in Patent Office: 1,096 Days
• Field of Search: 382/103, 382/107, 382/173, 382/236, 382/220, 348/699, 348/700, 348/169
• US Class Current: 382/103
• CPC Class Codes: G01S 3/7865 using correlation of the live video image with a stored image
• CPC Class Codes: G06T 7/246 using feature-based methods, e.g. the tracking of corners or segments
• CPC Class Codes: G06T 7/277 involving stochastic approaches, e.g. using Kalman filters
• CPC Class Codes: G06V 10/24 Aligning, centring, orientation detection or correction of the image
• CPC Class Codes: G06V 40/173 face re-identification, e.g. recognising unknown faces across different face tracks
• CPC Class Codes: G08B 13/19602 Image analysis to detect motion of the intruder, e.g. by frame subtraction
• CPC Class Codes: G08B 13/19608 Tracking movement of a target, e.g. by detecting an object predefined as a target, using target direction and or velocity to predict its new position
• CPC Class Codes: G08B 13/19673 Addition of time stamp, i.e. time metadata, to video stream

Abstract

A tracking system and method for evaluating whether image information for a region cluster of a video frame of a scene represents an hypothesis of an object to be tracked, such as a person. At least one real-world feature of a region cluster corresponding to an object to be tracked is generated. For example, the at least feature is at least one possible location of a predetermined portion of an object represented by the region cluster is determined based on a viewing angle of the scene of the video camera. A distance from the video camera to the object corresponding to the region is determined in real-world coordinates for each possible location of the predetermined portion of the region cluster. Real-world size and location information for the region cluster is determined based on the distance. The real-world size and location information for the region cluster is compared with statistical information for the particular type of object to determine a confidence value representing a measure of confidence that the region cluster represents the particular type of object.

Claims

1. A method for evaluating region cluster image information obtained from a scene to determine whether the region cluster represents an object to be tracked, the method comprising steps of:

generating at least one real-world feature for each region cluster that represents a characteristic of the region cluster in a real-world space comprising steps of;

(i) determining at least one possible location of a predetermined portion of the region cluster based on a viewing angle of the scene by a video camera which views the scene;

(ii) determining a distance from the video camera to the object corresponding to the region cluster in real-world coordinates for each possible location of the predetermined portion determined for the region cluster; and

(iii) generating real-world size and location information for the region cluster based on the distance; and

comparing the real-world size and location information for the region cluster with statistical information to generate a confidence value for each region cluster that represents a likelihood that the region cluster represents an object to be tracked.

2. The method of claim 1, wherein the predetermined portion is the feet of a person, and wherein the step of determining possible locations of the predetermined portion comprises steps of determining that the feet are on the bottom of the region cluster if the viewing angle of the video camera is oblique;

and if the viewing angle of the video camera is not oblique, then determining that the feet are located in at least two possible locations relative to a center of the view of the scene.

3. The method of claim 2, wherein if the viewing angle of the video camera is non-oblique, then determining that the feet are located on the bottom of the image region cluster if the image region is above the center of the view;

at the bottom of the region cluster or to the left of the region cluster if the region cluster is above and right of the center of the view, and at the top of the region cluster or on the left side of the region cluster if the image region is below and left of the center of the view, and at a centroid of the image region if the region cluster is within a predetermined radius of the center of the view.

4. The method of claim 1, wherein the step of determining the distance from the video camera to the object corresponding to the region cluster in real-world coordinates for each location of the predetermined portion comprises the step of determining a distance from a center of the region cluster to a Y-coordinate of the possible locations of the feet.

5. The method of claim 1, wherein the step of generating real-world size and location information comprises a step of computing a height, width and X-coordinate in real-world scale of an object in the scene corresponding to the region cluster.

6. The method of claim 1, wherein the step of determining the correspondence comprises determining a correspondence for those region clusters determined to have at least a minimum confidence value.

10. The system of claim 1, wherein the processor determines the distance from the video camera to the object corresponding to the region cluster in real-world coordinates for each location of the predetermined portion by determining a distance from a center of the region cluster to a Y-coordinate of the possible locations of the feet.

7. A system for evaluating region cluster image information obtained from a scene to determine whether the region cluster represents an object to be tracked, comprising:

(a) a video camera for viewing a scene;

(b) a frame grabber coupled to the video camera for generating video frames representing image information for the scene;

(c) a processor coupled to the frame grabber, the processor being programmed to;

generate at least one real-world feature for each region cluster that represents a characteristic of the region cluster in a real-world space by (i) determining at least one possible location of a predetermined portion of the region cluster based on a viewing angle of the scene by the video camera;

(ii) determining a distance from the video camera to the object corresponding to the region cluster in real-world coordinates for each possible location of the predetermined portion determined for the region cluster; and

(iii) generating real-world size and location information for the region cluster based on the distance; and

compare the real-world size and location information for the region cluster with statistical information to determine a confidence value for each region cluster that represents a likelihood that the region cluster represents an object to be tracked.

8. The system of claim 7, wherein the predetermined portion is the feet of a person, and wherein processor determines possible locations of the predetermined portion by determining that the feet are on the bottom of the region cluster if the viewing angle of the video camera is oblique;

and if the viewing angle of the video camera is not oblique, the processor determines that the feet are located in at least two possible locations relative to a center of the view of the scene.

9. The system of claim 8, wherein if the viewing angle of the video camera is non-oblique, then the processor determines that the feet are located on the bottom of the region cluster if the image region is above the center of the view;

at the bottom of the image region or to the left of the region cluster if the image region is above and right of the center of the view, and at the top of the region cluster or on the left side of the image region if the image region is below and left of the center of the view, and at a centroid of the image region if the region cluster is within a predetermined radius of the center of the view.

11. The system of claim 7, wherein the processor generates real-world size and location information by computing a height, width and X-coordinate in real-world scale of an object in the scene corresponding to the region cluster.

12. The system of claim 7, wherein processor determines the correspondence for those region clusters determined to have at least a minimum confidence value.