Method and System for Tracking Objects

• US 20150294158A1
• Filed: 04/10/2014
• Published: 10/15/2015
• Est. Priority Date: 04/10/2014
• Status: Active Grant

• Current Assignee: Disney Enterprises Incorporated (The Walt Disney Company)
• Original Assignee: Disney Enterprises Incorporated (The Walt Disney Company)
• Inventors: COLLINS, ROBERT T., CARR, GEORGE PETER
• Granted Patent: US 9,361,520 B2
• Time in Patent Office: Days
• US Class Current: 1/1
• CPC Class Codes: G06V 20/40 in video content extracting overlay text G06V20/62; video retrieval G06F16/70; processing of video elementary streams in video servers H04N21/234; processing of video elementary streams in video clients H04N21/44
• CPC Class Codes: G06V 20/42 of sport video content
• CPC Class Codes: G06V 20/52 Surveillance or monitoring of activities, e.g. for recognising suspicious objects recognising microscopic objects G06V20/69

Abstract

There is provided a system for tracking objects. The system includes a processor and a memory for storing a plurality of sensory data frames. The processor determines a first hypothesized location for each of the objects in each of the plurality of sensory data frames. For each of the plurality of sensory data frames, the processor determines probabilities that the first hypothesized location of each of the objects in a sensory data frame of the plurality of sensory data frames is the same as the first hypothesized location of another object in an adjacent sensory data frame. The processor computes a first optimal trajectory for each of the objects using an algorithm based on the probabilities, checks the first optimal trajectory for each of the objects, and accepts or rejects the first optimal trajectory for each of the objects.

Claims

1. A system for tracking objects, the system comprising:

a memory for storing a plurality of sensory data frames; and

a processor configured to;

determine a first hypothesized location for each of the objects in each of the plurality of sensory data frames;

for each of the plurality of sensory data frames, determine probabilities that the first hypothesized location of each of the objects in a sensory data frame of the plurality of sensory data frames is the same as the first hypothesized location of another object in an adjacent sensory data frame;

compute a first optimal trajectory for each of the objects using an algorithm based on the probabilities;

check the first optimal trajectory for each of the objects; and

accept or reject the first optimal trajectory for each of the objects based on the checking.

2. The system of claim 1, wherein the processor is further configured to:

determine a second hypothesized location for each of objects in each of the plurality of sensory data frames;

for each of the plurality of sensory data frames, determine second probabilities using the second hypothesized location for each of the objects;

compute a second optimal trajectory for each of the objects using the algorithm based on the second probabilities;

check the second optimal trajectory for each of objects; and

accept or reject the second optimal trajectory for each of the objects based on the checking.

4. The system of claim 1, wherein the checking the first optimal trajectory for each of the objects includes at least one of checking if the first optimal trajectory is probable using occupancy maps and checking appearances of each of the objects during a duration of the first optimal trajectory.

5. The system of claim 1, wherein the plurality of sensory data frames correspond to a plurality of video frames.

6. The system of claim 5, wherein determining the first hypothesized location of the objects in each of the plurality of video frames includes:

performing a background subtraction on the video to obtain a background reference frame;

comparing each of the plurality of video frames to the background reference frame to obtain a foreground mask for each of the plurality of video frames; and

generating an occupancy map for each of the plurality of video frames using the foreground mask for each of the plurality of video frames.

7. The system of claim 6, wherein the first hypothesized location for each of the objects corresponds to a peak in the occupancy map for each of the plurality of video frames.

8. The system of claim 1, wherein the algorithm includes at least one of a maximum flow algorithm or a Munkres-Kuhn (Hungarian) algorithm.

9. The system of claim 1, wherein each of the objects includes a person.

10. The system of claim 1, wherein the processor is further configured to:

locate each of the objects in the plurality of sensory data frames using the first hypothesized location of each of the objects; and

mark each of the objects in the plurality of sensory data frames using a detection box.

3. The system of claim 3, wherein determining the second hypothesized location for each of the objects includes at least one of changing the first hypothesized location of at least one of the objects to get the second hypothesized location of the at least one of the objects in the plurality of sensory data frames, adding an object to the objects, or removing an object from the objects.

11. A method for tracking objects, the method comprising:

obtaining a plurality of sensory data frames;

determining a first hypothesized location for each of the objects in each of the plurality of sensory data frames;

for each of the plurality of sensory data frames, determining probabilities that the first hypothesized location of each of the objects in a sensory data frame of the plurality of sensory data frames is the same as the first hypothesized location of another object in an adjacent sensory data frame;

computing a first optimal trajectory for each of the objects using an algorithm based on the probabilities;

checking the first optimal trajectory for each of the objects; and

accepting or rejecting the first optimal trajectory for each of the objects based on the checking.

12. The method of claim 11, further comprising:

determining a second hypothesized location for each of objects in each of the plurality of sensory data frame;

for each of the plurality of data sensory frames, determining second probabilities using the second hypothesized location for each of the objects;

computing a second optimal trajectory for each of the objects using the algorithm based on the second probabilities;

checking the second optimal trajectory for each of objects; and

accepting or rejected the second optimal trajectory for each of the objects based on the checking.

13. The method of claim 12, wherein determining the second hypothesized location for each of the objects includes at least one of changing the first hypothesized location of at least one of the objects to get the second hypothesized location of the at least one of the objects in the plurality of sensory data frames, adding an object to the objects, or removing an object from the objects.

14. The method of claim 11, wherein checking the first optimal trajectory for each of the objects includes at least one of checking if the first optimal trajectory is probable using occupancy maps and checking appearances of each of the objects during a duration of the first optimal trajectory.

15. The method of claim 11, wherein the plurality of sensory data frames correspond to a plurality of video frames.

16. The method of claim 15, wherein determining the first hypothesized location of the objects in each of the plurality of video frames includes:

performing a background subtraction on the video to obtain a background reference frame;

comparing each of the plurality of video frames to the background reference frame to obtain a foreground mask for each of the plurality of video frames; and

generating an occupancy map for each of the plurality of video frames using the foreground mask for each of the plurality of video frames.

17. The method of claim 16, wherein the first hypothesized location for each of the objects corresponds to a peak in the occupancy map for each of the plurality of video frames.

18. The method of claim 11, wherein the algorithm includes at least one of a maximum flow algorithm or a Munkres-Kuhn (Hungarian) algorithm.

19. The method of claim 11, wherein each of the objects includes a person.

20. The method of claim 11, further comprising:

locating the objects in each of the plurality of sensory data frames using the first hypothesized location for each of the objects; and

marking each of the objects in the plurality of sensory data frames using a detection box.