Field calibration of stereo cameras with a projector

Patent metadata

Field	Value
Publication number	US-2018249142-A1
Application number	US-201715443709-A
Country	US
Kind code	A1
Filing date	Feb 27, 2017
Priority date	Feb 27, 2017
Publication date	Aug 30, 2018
Grant date	—

Abstract

Calibration in the field is described for stereo and other depth camera configurations using a projector One example includes imaging the first and the second feature in a first camera of the camera system wherein the distance from the first camera to the projector is known, imaging the first and the second feature in a second camera of the camera system, wherein the distance from the second camera to the projector is known, determining a first disparity between the first camera and the second camera to the first feature, determining a second disparity between the first camera and the second camera to the second feature, and determining an epipolar alignment error of the first camera using the first and the second disparities.

First claim

What is claimed is:

1 . A method comprising:
projecting a first and a second feature onto a surface from a projector of a camera system; imaging the first and the second feature in a first camera of the camera system wherein the distance from the first camera to the projector is known; imaging the first and the second feature in a second camera of the camera system, wherein the distance from the second camera to the projector is known; determining a first disparity between the first camera and the second camera to the first feature; determining a second disparity between the first camera and the second camera to the second feature; and determining an epipolar alignment error of the first camera using the first and the second disparities.

2 . The method of claim 1 , wherein the distance from the camera system to the first feature and the distance from the camera system to the second feature is not known.

3 . The method of claim 2 , wherein the distance from the camera system to the first feature is different from the distance from the camera system to the second feature.

4 . The method of claim 1 , wherein the epipolar alignment error comprises a pointing error.

5 . The method of claim 1 , further comprising:
determining a third disparity between the projector and the first camera; determining a fourth disparity between the projector and the second camera; and determining an epipolar alignment error of the second camera using the first, second, third, and fourth disparities.

6 . The method of claim 1 , wherein the camera system has a planar base and the projector, first camera, and second camera are all mounted to the base.

7 . The method of claim 1 , wherein the projector projects a patterned light onto the surface, wherein the pattern defines the first and the second features.

8 . The method of claim 1 , further comprising correcting for magnification errors and pointing errors that are orthogonal to the stereo axis before determining an epipolar alignment error.

9 . The method of claim 1 , further comprising:
generating calibration parameters using the epipolar alignment error; and warping images of the first and the second camera using the calibration parameters.

10 . The method of claim 1 , wherein determining an epipolar alignment error comprises for the first feature:
extending a first epipolar line from the imaged first feature from the first camera; extending a second epipolar line from the imaged first feature from the second camera; determining an orthogonal distance from the projected first feature to the first epipolar line; determining an orthogonal distance from the projected first feature to the second epipolar line; and determining scale and displacement parameters using the orthogonal distances.

11 . The method of claim 1 , wherein determining an epipolar alignment error comprises minimizing a rectification error.

12 . The method of claim 11 , wherein minimizing a rectification error comprises applying a matrix multiplication form to known parameters.

13 . The method of claim 1 , further comprising:
projecting additional features onto the surface from the projector; imaging the additional features in the first and the second camera; and determining additional disparities each based on an additional feature, wherein determining an epipolar alignment error further comprises determining an epipolar alignment error using the determined additional disparities.

14 . A machine-readable medium having instructions that when operated on by the machine cause the machine to perform operations comprising:
receiving a first image captured by a first camera including a first and a second feature that are projected from a projector, wherein the distance from the first camera to the projector is known; receiving a second image captured by a second camera including the first and the second feature that are projected from the projector, wherein the distance from the second camera to the projector and the distance from the second camera to the first camera are known; determining a first disparity between the first camera and the second camera to the first feature; determining a second disparity between the first camera and the second camera to the second feature; and determining an epipolar alignment error of the first camera using the first and the second disparities.

15 . The medium of claim 14 , the operations further comprising correcting for magnification errors and pointing errors that are orthogonal to a stereo axis between the first camera and the second camera before determining an epipolar alignment error.

16 . The medium of claim 14 , wherein determining an epipolar alignment error comprises for the first feature:
extending a first epipolar line from the imaged first feature from the first camera; extending a second epipolar line from the imaged first feature from the second camera; determining an orthogonal distance from the projected first feature to the first epipolar line; determining an orthogonal distance from the projected first feature to the second epipolar line; and determining scale and displacement parameters using the orthogonal distances.

17 . The medium of claim 14 , wherein determining an epipolar alignment error comprises minimizing a rectification error by applying a matrix multiplication form to known parameters.

18 . A camera system comprising:
a projector of the camera system to project a first and a second feature onto a surface from a projector of a camera system; a first image sensor of the camera system to image the first and the second feature wherein the distance from the first image sensor to the projector is known; a second camera of the camera system to image the first and the second feature, wherein the distance from the second image sensor to the projector is known; and a processor to determine a first disparity between the first image sensor and the second image sensor to the first feature, to determine a second disparity between the first image sensor and the second image sensor to the second feature, and to determine an epipolar alignment error of the first image sensor using the first and the second disparities.

19 . The system of claim 18 , having a planar base and wherein the projector, first camera, and second camera are all mounted to the base.

20 . The system of claim 18 , further comprising a mass memory to store images and wherein the images of the first and second feature from the first and the second image sensor are stored in the memory for post processing by the processor.

Assignees

• Intel Corp

Inventors

• Hicks Richmond
• Raniwala Sundeep

Classifications

• G01B11/2504

  Calibration devices · CPC title

• H04N17/002 Primary

  for television cameras · CPC title

• H04N13/128

  Adjusting depth or disparity · CPC title

• G06T2207/10048

  Infrared image · CPC title

• G06T2207/10024

  Color image · CPC title