Homography error correction using sensor locations

Patent metadata

Field	Value
Publication number	US-11301691-B2
Application number	US-202017105230-A
Country	US
Kind code	B2
Filing date	Nov 25, 2020
Priority date	Oct 25, 2019
Publication date	Apr 12, 2022
Grant date	Apr 12, 2022

Abstract

An object tracking system that includes a first sensor and a second sensor that are each configured to capture frames of at least a portion of a global plane for a space. The system is configured to identify a pixel location for a marker within a frame from the first sensor and to determine an (x,y) coordinate for the marker using a first homography. The system is further configured to identify a pixel location for a different marker in a frame from the second sensor and to determine an (x,y) coordinate for the marker using a second homography. The system is further configured to determine a distance difference between the computed distance between the (x,y) coordinates and an actual distance. The system is further configured to recompute the first homography and/or the second homography in response to determining that the distance difference exceeds a difference threshold level.

First claim

The invention claimed is:

1. An object tracking system, comprising:
a first sensor configured to capture frames of a global plane for at least a first portion of a space, wherein:
the global plane represents (x,y) coordinates for the at least a portion of the space; and
each frame comprises a plurality of pixels;

a second sensor configured to capture frames of the global plane for at least a second portion of the space; and
a tracking system operably coupled to the first sensor and the second sensor, comprising:
one or more memories operable to store:
a first homography associated with the first sensor, wherein the first homography comprises coefficients that translate between pixel locations in a frame from the first sensor and (x,y) coordinates in the global plane; and
a second homography associated with the second sensor, wherein the second homography comprises coefficients that translate between pixel locations in a frame from the second sensor and (x,y) coordinates in the global plane; and

one or more processors operably coupled to the one or more memories, configured to:
receive a first frame from the first sensor;
identify a first pixel location for a first marker within the first frame;
determine a first (x,y) coordinate for the first marker in the global plane by applying the first homography to the first pixel location;
receive a second frame from the second sensor;
identify a second pixel location for a second marker in the second frame;
determine a second (x,y) coordinate for the second marker in the global plane by applying the second homography to the second pixel location;
determine a computed distance between the first (x,y) coordinate and the second (x,y) coordinate;
receive an actual distance between the first marker and the second marker;
determine a distance difference between the computed distance and the actual distance;
compare the distance difference to a difference threshold level; and
recompute at least one of the first homography and the second homography in response to determining that the distance difference exceeds the difference threshold level.

2. The system of claim 1 , wherein the first sensor and the second sensor are each further configured to capture depth information.

3. The system of claim 1 , wherein:
identifying the first pixel location within the first frame comprises detecting the first marker within the first frame; and
the first pixel location corresponds with a location of the first marker within the first frame.

4. The system of claim 1 , wherein:
the first pixel location corresponds with a first pixel in a center of the first frame; and
the second pixel location corresponds with a second pixel in a center of the second frame.

5. The system of claim 1 , wherein:
the first marker is a first light source that is detectable by the first sensor; and
the second marker is a second light source that is detectable by the second sensor.

6. The system of claim 1 , wherein the global plane is parallel with a floor of the space.

7. The system of claim 1 , wherein recomputing the homography comprises:
receiving a third (x,y) coordinate identifying a first x-value and a first y-value in the global plane where a third marker is located in the space, wherein the third marker is a first object identifying a first location in the space;
receiving a fourth (x,y) coordinate identifying a second x-value and a second y-value in the global plane where a fourth marker is located in the space, wherein the fourth marker is a second object identifying a second location in the space;
receiving a new frame from the first sensor;
identifying the third marker and the fourth marker within the new frame;
determining a third pixel location in the new frame for the third marker, wherein the third pixel location comprises a first pixel row and a first pixel column of the new frame;
determining a fourth pixel location in the second frame for the fourth marker, wherein the fourth pixel location comprises a second pixel row and a second pixel column of the new frame; and
recomputing the first homography based on the third (x,y) coordinate, the fourth (x,y) coordinate, the third pixel location, and the fourth pixel location.

8. A homography error correction method, comprising:
receiving a first frame from a first sensor, wherein the first sensor configured to capture frames of a global plane for at least a first portion of a space, wherein:
the global plane represents (x,y) coordinates for the at least a portion of the space; and
each frame comprises a plurality of pixels;

identifying a first pixel location for a first marker within the first frame;
determining a first (x,y) coordinate for the first marker in the global plane by applying a first homography to the first pixel location, wherein:
the first homography is associated with the first sensor; and
the first homography comprises coefficients that translate between pixel locations in a frame from the first sensor and (x,y) coordinates in the global plane;

receiving a second frame from a second sensor, wherein the second sensor configured to capture frames of the global plane for at least a second portion of the space;
identifying a second pixel location for a second marker in the second frame;
determining a second (x,y) coordinate for the second marker in the global plane by applying a second homography to the second pixel location, wherein:
the second homography is associated with the second sensor; and
the second homography comprises coefficients that translate between pixel locations in a frame from the second sensor and (x,y) coordinates in the global plane;

determining a computed distance between the first (x,y) coordinate and the second (x,y) coordinate;
receiving an actual distance between the first marker and the second marker;
determining a distance difference between the computed distance and the actual distance;
comparing the distance difference to a difference threshold level; and
recomputing at least one of the first homography and the second homography in response to determining that the distance difference exceeds the difference threshold level.

9. The method of claim 8 , wherein the first sensor and the second sensor are each further configured to capture depth information.

10. The method of claim 8 , wherein:
identifying the first pixel location within the first frame comprises detecting the first marker within the first frame; and
the first pixel location corresponds with a location of the first marker within the first frame.

11. The method of claim 8 , wherein:
the first pixel location corresponds with a first pixel in a center of the first frame; and
the second pixel location corresponds with a second pixel in a center of the second frame.

12. The method of claim 8 , wherein:
the first marker is a first light source that is detectable by the first sensor; and
the second marker is a second light source that is detectable by the second sensor.

13. The method of claim 8 , wherein the global plane is parallel with a floor of the space.

14. The method of claim 8 , wherein recomputing the homography comprises:
receiving a third (x,y) coordinate identifying a first x-value and a first y-value in the global plane where a third marker is located in the space, wherein the third marker is a first object identifying a first location in the space;
receiving a fourth (x,y) coordinate identifying a second x-value and a second y-value in the global plane where a fourth marker is located in the space, wherein the fourth marker is a second object identifying a second location

Assignees

• 7 Eleven Inc

Inventors

• Mirza Shahmeer Ali
• Krishnamurthy Sailesh Bharathwaaj
• Dalal Kyle

Classifications

• G06V40/20 Primary

  Movements or behaviour, e.g. gesture recognition (recognition of facial expressions G06V40/16) · CPC title

• G06V20/52 Primary

  Surveillance or monitoring of activities, e.g. for recognising suspicious objects (recognising microscopic objects G06V20/69) · CPC title

• G06V10/82

  using neural networks · CPC title

• G06V10/80

  Fusion, i.e. combining data from various sources at the sensor level, preprocessing level, feature extraction level or classification level (multimodal speaker identification or verification G10L17/10) · CPC title

• G06T2207/30232

  Surveillance · CPC title