Tracking positions using a scalable position tracking system

1. A system, comprising:

an array of cameras positioned above a space, each camera of the array of cameras configured to capture a video of a portion of the space, the space containing a person;

a first camera client configured to:

receive a first plurality of frames of a first video from a first camera of the array of cameras, each frame of the first plurality of frames showing the person within the space; and

for each frame of the first plurality of frames:

determine a bounding area around the person shown in that frame; and

generate a timestamp of when that frame was received by the first camera client;

a second camera client configured to:

receive a second plurality of frames of a second video from a second camera of the array of cameras, each frame of the second plurality of frames showing the person within the space; and

for each frame of the second plurality of frames:

determine a bounding area around the person shown in that frame; and

generate a timestamp of when that frame was received by the second camera client; and

a camera server separate from the first and second camera clients, the camera server configured to:

for each frame of the first plurality of frames, assign, based at least on the timestamp of when that frame was received by the first camera client, coordinates defining the bounding area around the person shown in that frame to one of a plurality of time windows;

for each frame of the second plurality of frames, assign, based at least on the timestamp of when that frame was received by the second camera client, coordinates defining the bounding area around the person shown in that frame to one of the plurality of time windows;

process the coordinates assigned to the first time window by:

calculating, based at least on the coordinates that (1) define bounding areas around the person shown in the first plurality of frames and (2) are assigned to the first time window, a combined coordinate for the person during the first time window for the first video from the first camera; and

calculating, based at least on the coordinates that (1) define bounding areas around the person shown in the second plurality of frames and (2) are assigned to the first time window, a combined coordinate for the person during the first time window for the second video from the second camera.

2. The system of claim 1 , wherein the camera server is further configured to determine, based at least on the combined coordinate for the person during the first time window for the first video from the first camera and the combined coordinate for the person during the first time window for the second video from the second camera, a position of the person within the space during the first time window.

3. The system of claim 1 , wherein the combined coordinate for the person during the first time window for the first video from the first camera comprises an average of the coordinates that (1) define bounding areas around the person shown in the first plurality of frames and (2) are assigned to the first time window.

4. The system of claim 1 , wherein processing the coordinates assigned to the first time window is performed in response to:

determining that coordinates assigned to the first time window comprise coordinates from frames from a number of cameras in the array of cameras; and

determining that the number of cameras exceeds a threshold.

5. The system of claim 1 , wherein processing the coordinates assigned to the first time window is performed in response to:

determining that coordinates assigned to the second time window comprise coordinates from frames from a number of cameras in the array of cameras; and

determining that the number of cameras exceeds a threshold.

6. The system of claim 1 , wherein processing the coordinates assigned to the first time window is performed in response to determining that coordinates assigned to the first time window have not been processed for a period of time exceeding a timeout.

7. The system of claim 6 , the camera server is further configured to:

in response to determining that coordinates assigned to first time window have not been processed for the period of time exceeding the timeout:

determine that coordinates assigned to the second time window comprise coordinates from frames from a first number of cameras in the array of cameras; and

lower a threshold to the first number;

determine that coordinates assigned to a third time window comprise coordinates from frames from a second number of cameras in the array of cameras;

determine that the second number exceeds the first number; and

in response to determining that the second number exceeds the first number, increase the threshold to the second number.

8. The system of claim 1 , wherein determining that coordinates assigned to the first time window should be processed comprises determining that a frame has been received from each camera in the array of cameras during the second time window.

9. The system of claim 1 , wherein the first camera client is further configured to communicate the coordinates defining bounding areas around the person shown in the first plurality of frames and the coordinates defining bounding areas around the person shown in the second plurality of frames as a batch to the camera server in response to determining that during the first time window, the first camera client has received a frame from each camera in the array of cameras that is communicatively coupled to the first camera client.

10. The system of claim 1 , wherein:

the first camera client implements a first clock used to generate the timestamps of when each frame of the first plurality of frames was received by the first camera client;

the second camera client implements a second clock used to generate the timestamps of when each frame of the second plurality of frames was received by the second camera client; and

the camera server implements a third clock, the first, second, and third clocks are synchronized using a clock synchronization protocol.

11. The system of claim 1 , wherein the array of cameras is arranged in a grid such that:

each camera that is communicatively coupled to the first camera client is not directly adjacent in the grid to another camera that is communicatively coupled to the first camera client; and

a camera that is communicatively coupled to the first camera client is diagonal in the grid to another camera that is communicatively coupled to the first camera client.

12. The system of claim 1 , wherein:

the space further contains a second person, each frame of the first and second plurality of frames showing the second person;

the first camera client is further configured to, for each frame of the first plurality of frames, determine a bounding area around the second person shown in that frame;

a second camera client configured to, for each frame of the second plurality of frames, determine a bounding area around the second person shown in that frame; and

the camera server is further configured to:

for each frame of the first plurality of frames, assign, based at least on the timestamp of when that frame was received by the first camera client, coordinates defining the bounding area around the second person shown in that frame to one of a plurality of time windows;

for each frame of the second plurality of frames, assign, based at least on the timestamp of when that frame was received by the second camera client, coordinates defining the bounding area around the second person shown in that frame to one of the plurality of time windows;

further process the coordinates assigned to the first time window by:

calculating, based at least on the coordinates that (1) define bounding areas around the second person shown in the first plurality of frames and (2) are assigned to the first time window, a combined coordinate for the second person during the first time window for the first video from the first camera; and

calculating, based at least on the coordinates that (1) define bounding areas around the second person shown in the second plurality of frames and (2) are assigned to the first time window, a combined coordinate for the second person during the first time window for the second video from the second camera.

13. A method comprising:

receiving, by a first camera client, a first plurality of frames of a first video from a first camera of an array of cameras, the array of cameras positioned above a space, each camera of the array of cameras configured to capture a video of a portion of the space, the space containing a person, each frame of the first plurality of frames showing the person within the space;

for each frame of the first plurality of frames:

determining, by the first camera client, a bounding area around the person shown in that frame; and

generating, by the first camera client, a timestamp of when that frame was received by the first camera client;

receiving, by a second camera client, a second plurality of frames of a second video from a second camera of the array of cameras, each frame of the second plurality of frames showing the person within the space;

for each frame of the second plurality of frames:

determining, by the second camera client, a bounding area around the person shown in that frame; and

generating, by the second camera client, a timestamp of when that frame was received by the second camera client; and

for each frame of the first plurality of frames, assigning by a camera server separate from the first and second camera clients, based at least on the timestamp of when that frame was received by the first camera client, coordinates defining the bounding area around the person shown in that frame to one of a plurality of time windows;

for each frame of the second plurality of frames, assigning by the camera server, based at least on the timestamp of when that frame was received by the second camera client, coordinates defining the bounding area around the person shown in that frame to one of the plurality of time windows;

processing, by the camera server, the coordinates assigned to the first time window by:

calculating, based at least on the coordinates that (1) define bounding areas around the person shown in the first plurality of frames and (2) are assigned to the first time window, a combined coordinate for the person during the first time window for the first video from the first camera; and

calculating, based at least on the coordinates that (1) define bounding areas around the person shown in the second plurality of frames and (2) are assigned to the first time window, a combined coordinate for the person during the first time window for the second video from the second camera.

14. The method of claim 13 , further comprising determining, by the camera server, based at least on the combined coordinate for the person during the first time window for the first video from the first camera and the combined coordinate for the person during the first time window for the second video from the second camera, a position of the person within the space during the first time window.

15. The method of claim 13 , wherein the combined coordinate for the person during the first time window for the first video from the first camera comprises an average of the coordinates that (1) define bounding areas around the person shown in the first plurality of frames and (2) are assigned to the first time window.

16. The method of claim 13 , wherein processing the coordinates assigned to the first time window is performed in response to:

determining that coordinates assigned to the first time window comprise coordinates from frames from a number of cameras in the array of cameras; and

determining that the number of cameras exceeds a threshold.

17. The method of claim 13 , wherein processing the coordinates assigned to the first time window is performed in response to:

determining that coordinates assigned to the second time window comprise coordinates from frames from a number of cameras in the array of cameras; and

determining that the number of cameras exceeds a threshold.

18. The method of claim 13 , wherein processing the coordinates assigned to the first time window is performed in response to determining that coordinates assigned to the first time window have not been processed for a period of time exceeding a timeout.

19. The method of claim 18 , further comprising:

in response to determining that coordinates assigned to first time window have not been processed for the period of time exceeding the timeout:

determining that coordinates assigned to the second time window comprise coordinates from frames from a first number of cameras in the array of cameras; and

lowering a threshold to the first number;

determining that coordinates assigned to a third time window comprise coordinates from frames from a second number of cameras in the array of cameras;

determining that the second number exceeds the first number; and

in response to determining that the second number exceeds the first number, increasing the threshold to the second number.

20. The method of claim 13 , wherein determining that coordinates assigned to the first time window should be processed comprises determining that a frame has been received from each camera in the array of cameras during the second time window.

21. The method of claim 13 , further comprising communicating, by the first camera client, the coordinates defining bounding areas around the person shown in the first plurality of frames and the coordinates defining bounding areas around the person shown in the second plurality of frames as a batch to the camera server in response to determining that during the first time window, the first camera client has received a frame from each camera in the array of cameras that is communicatively coupled to the first camera client.

22. The method of claim 13 , further comprising:

implementing, by the first camera client, a first clock used to generate the timestamps of when each frame of the first plurality of frames was received by the first camera client;

implementing, by the second camera client, a second clock used to generate the timestamps of when each frame of the second plurality of frames was received by the second camera client; and

implementing, by the camera server a third clock, the first, second, and third clocks are synchronized using a clock synchronization protocol.

23. The method of claim 13 , wherein the array of cameras is arranged in a grid such that:

each camera that is communicatively coupled to the first camera client is not directly adjacent in the grid to another camera that is communicatively coupled to the first camera client; and

a camera that is communicatively coupled to the first camera client is diagonal in the grid to another camera that is communicatively coupled to the first camera client.

24. The method of claim 13 , wherein the space further contains a second person, each frame of the first and second plurality of frames showing the second person, and the method further comprising:

for each frame of the first plurality of frames, determining, by the first camera client, a bounding area around the second person shown in that frame;

for each frame of the second plurality of frames, determining, by the second camera client, a bounding area around the second person shown in that frame;

for each frame of the first plurality of frames, assigning, by the camera server, based at least on the timestamp of when that frame was received by the first camera client, coordinates defining the bounding area around the second person shown in that frame to one of a plurality of time windows;

for each frame of the second plurality of frames, assigning, by the camera server, based at least on the timestamp of when that frame was received by the second camera client, coordinates defining the bounding area around the second person shown in that frame to one of the plurality of time windows;

further processing by the camera server, the coordinates assigned to the first time window by:

calculating, based at least on the coordinates that (1) define bounding areas around the second person shown in the first plurality of frames and (2) are assigned to the first time window, a combined coordinate for the second person during the first time window for the first video from the first camera; and

calculating, based at least on the coordinates that (1) define bounding areas around the second person shown in the second plurality of frames and (2) are assigned to the first time window, a combined coordinate for the second person during the first time window for the second video from the second camera.

25. A system, comprising:

an array of cameras positioned above a space, each camera of the array of cameras configured to capture a video of a portion of the space, the space containing an object;

a first camera client configured to:

receive a first plurality of frames of a first video from a first camera of the array of cameras, each frame of the first plurality of frames showing the object within the space; and

for each frame of the first plurality of frames:

determine a bounding area around the object shown in that frame; and

generate a timestamp of when that frame was received by the first camera client;

a second camera client configured to:

receive a second plurality of frames of a second video from a second camera of the array of cameras, each frame of the second plurality of frames showing the object within the space; and

for each frame of the second plurality of frames:

determine a bounding area around the object shown in that frame; and

generate a timestamp of when that frame was received by the second camera client; and

a camera server separate from the first and second camera clients, the camera server configured to:

for each frame of the first plurality of frames, assign, based at least on the timestamp of when that frame was received by the first camera client, coordinates defining the bounding area around the object shown in that frame to one of a plurality of time windows;

for each frame of the second plurality of frames, assign, based at least on the timestamp of when that frame was received by the second camera client, coordinates defining the bounding area around the object shown in that frame to one of the plurality of time windows;

process the coordinates assigned to the first time window by:

calculating, based at least on the coordinates that (1) define bounding areas around the object shown in the first plurality of frames and (2) are assigned to the first time window, a combined coordinate for the object during the first time window for the first video from the first camera; and

calculating, based at least on the coordinates that (1) define bounding areas around the object shown in the second plurality of frames and (2) are assigned to the first time window, a combined coordinate for the object during the first time window for the second video from the second camera.