Estimating heading misalignment between a device and a person using optical sensor

Patent metadata

Field	Value
Publication number	US-9818037-B2
Application number	US-201615014957-A
Country	US
Kind code	B2
Filing date	Feb 3, 2016
Priority date	Feb 4, 2015
Publication date	Nov 14, 2017
Grant date	Nov 14, 2017

Abstract

This disclosure is about a method and apparatus for estimating heading misalignment between a device and a person, where the device/apparatus comprises an optical sensor, the optical sensor capable of capturing an image of the person. At least one physical feature of the person is extracted from the image. A misalignment angle is estimated using the at least one physical feature of the person extracted from the image.

First claim

What is claimed is:

1. A method for estimating heading misalignment between a device and a person, wherein the device comprises an optical sensor capable of capturing an image, the method comprising:
capturing an image of the person by the optical sensor;
extracting at least one physical feature of the person from the image; and
estimating a fine misalignment angle between the device and the person using the at least one physical feature of the person extracted from the image, the fine misalignment angle providing a value of the heading misalignment between the device and the person, wherein the heading misalignment is a difference between a heading of the device and a direction of motion of the person.

2. The method as recited in claim 1 , wherein the extracting the at least one physical feature of the person from the image of the person further comprises:
performing preprocessing on the image.

3. The method as recited in claim 1 , wherein the extracting the at least one physical feature of the person from the image of the person comprises:
performing edge detection on the image of the person; and
identifying the at least one physical feature of the person using detected edges of the image.

4. The method as recited in claim 3 , wherein the extracting the at least one physical feature of the person from the image of the person further comprises:
performing a Hough transform using detected edges of the image.

5. The method as recited in claim 4 , wherein the Hough transform is used to detect at least one or a combination of the following: (i) straight lines; and (ii) circles.

6. The method as recited in claim 1 , wherein the extracting the at least one physical feature of the person from the image of the person comprises:
performing color analysis on the image of the person;
detecting skin color of the person; and
identifying the at least one physical feature of the person using the skin color.

7. The method as recited in claim 1 , wherein the at least one physical feature of the person comprises at least one or a combination of a head, shoulders, ears, eyes, a nose, a neck, a mouth, and eyeglasses.

8. The method as recited in claim 1 , further comprising:
estimating a coarse misalignment between the device and the person, the coarse misalignment providing an indication of relative orientation of the device from an orientation of the image relative to the device.

9. The method as recited in claim 8 , wherein the estimating a coarse misalignment between the device and the person comprises:
determining a relation between the at least one physical feature of the person relative to a frame of the image; and
identifying the coarse misalignment based on the relation.

10. The method as recited in claim 8 , wherein the coarse misalignment comprises one of the following: portrait with zero degrees, landscape with 90 degrees, and landscape with −90 degrees.

11. The method as recited in claim 1 , wherein the estimating a fine misalignment angle between the device and the person comprises:
determining an image length ratio of a first physical feature of the person to a second physical feature of the person, wherein an actual length ratio of the first physical feature of the person to a second physical feature of the person is known; and
determining the fine misalignment angle using the image length ratio and the actual length ratio.

12. The method as recited in claim 1 , wherein the estimating a fine misalignment angle between the device and the person comprises:
determining an image length ratio of a distance from a middle neck point to a left shoulder side to a distance from the middle neck point to a right shoulder side; and
determining the fine misalignment angle using the image length ratio.

13. The method as recited in claim 1 further comprising:
obtaining a heading of the person from a heading of the device and the fine misalignment angle.

14. The method as recited in claim 1 further comprising:
using the fine misalignment angle to provide an enhanced navigation solution.

15. A portable device comprising:
an optical sensor configured to capture images;
a processor configured to receive images from the optical sensor and operable to estimate heading misalignment between the portable device and a person, wherein the processor is operable to:
receive an image of the person from the optical sensor;
extract at least one physical feature of the person from the image; and
estimate a fine misalignment angle between the portable device and the person using the at least one physical feature of the person extracted from the image, the fine misalignment angle providing a value of the heading misalignment between the portable device and the person, wherein the heading misalignment is a difference between a heading of the portable device and a direction of motion of the person.

16. The portable device as recited in claim 15 , wherein the processor is further operable to:
perform preprocessing on the image.

17. The portable device as recited in claim 15 , wherein the processor is further operable to:
perform edge detection on the image of the person; and
identify the at least one physical feature of the person using detected edges of the image.

18. The portable device as recited in claim 17 , wherein the processor is further operable to:
perform a Hough transform using detected edges of the image.

19. The portable device as recited in claim 18 , wherein the Hough transform is used to detect at least one or a combination of the following: (i) straight lines; and (ii) circles.

20. The portable device as recited in claim 15 , wherein the processor is further operable to:
perform color analysis on the image of the person;
detect skin color of the person; and
identify the at least one physical feature of the person using the skin color.

21. The portable device as recited in claim 15 , wherein the at least one physical feature of the person comprises at least one or a combination of a head, shoulders, ears, eyes, a nose, a neck, a mouth, and eyeglasses.

22. The portable device as recited in claim 15 , wherein the processor is further operable to:
estimate a coarse misalignment between the portable device and the person, the coarse misalignment providing an indication of relative orientation of the portable device from an orientation of the image relative to the portable device.

23. The portable device as recited in claim 22 , wherein the processor is further operable to:
determine a relation between the at least one physical feature of the person relative to a frame of the image; and
identify the coarse misalignment based on the relation.

24. The portable device as recited in claim 22 , wherein the coarse misalignment comprises one of the following: portrait with zero degrees, landscape with 90 degrees, and landscape with −90 degrees.

25. The portable device as recited in claim 15 , wherein the processor is further operable to:
determine an image length ratio of a first physical feature of the person to a second physical feature of the person, wherein an actual length ratio of the first physical feature of the person to a second physical feature of the person is known; and
determine the fine misalignment angle using the image length ratio and the actual length ratio.

26. The portable device as recited in claim 15 , wherein the processor is further operabl

Assignees

• Invensense Inc

Inventors

• Al-Hamad Amr
• Ali Abdelrahman
• Georgy Jacques
• Goodall Christopher

Classifications

• G01C11/04 Primary

  Interpretation of pictures · CPC title

• G06V10/242

  by image rotation, e.g. by 90 degrees · CPC title

• G06V10/56

  relating to colour · CPC title

• G06V10/245

  by locating a pattern; Special marks for positioning · CPC title

• G06K9/4652

  Physics · mapped topic