Method and device for verification

Patent metadata

Field	Value
Publication number	US-10872439-B2
Application number	US-201816176652-A
Country	US
Kind code	B2
Filing date	Oct 31, 2018
Priority date	Nov 1, 2017
Publication date	Dec 22, 2020
Grant date	Dec 22, 2020

Abstract

The present invention provides a method and a device for verification applied to the field of image processing. The method includes: acquiring a common two-dimensional image, an infrared image and a depth map of a calibration object, wherein the calibration object carries a first pattern; calculating coordinates of an image formed on the infrared image by the first pattern, acquiring a depth value of the first pattern according to the depth map, and calculating a position of a projection formed on the common two-dimensional image by a three-dimensional point cloud corresponding to the image formed on the infrared image by the first pattern, and calculating a distance between an image and the projection formed on the common two-dimensional image by the first pattern. The method and device for verification provided by the present invention can verify a calibration parameter obtained after calibration performed by a depth camera device, enhancing calibration accuracy.

First claim

What is claimed is:

1. A method for verification, characterized by, the method comprising:
acquiring a common two-dimensional image, an infrared image and a depth map of a calibration object, wherein the calibration object carries a first pattern; and
(1) according to the infrared image, calculating coordinates of an image formed on the infrared image by the first pattern, (2) according to the depth map, acquiring a depth value of the first pattern, (3) according to a parameter obtained from calibration, the coordinates of the image formed on the infrared image by the first pattern, and the depth value of the first pattern, calculating a position of a projection formed on the common two-dimensional image by a three-dimensional point cloud corresponding to the image formed on the infrared image by the first pattern, and (4) calculating a distance between an image formed on the common two-dimensional image by the first pattern and the projection of the three-dimensional point cloud on the common two-dimensional image; or
(a) according to the common two-dimensional image, calculating coordinates of an image formed on the common two-dimensional image by the first pattern, (b) according to the depth map, acquiring a depth value of the first pattern, (c) calculating a position of a projection formed on the infrared image by a three-dimensional point cloud corresponding to the image formed on the common two-dimensional image by the first pattern, and (d) calculating a distance between an image formed on the infrared image by the first pattern and the projection of the three-dimensional cloud point on the infrared image.

2. The method according to claim 1 , characterized by, the method further comprising:
determining a verification result according to the distance and a first threshold.

3. The method according to claim 1 , characterized in that:
the calibration object comprises at least three equal-sized calibration boards, wherein the first patterns carried by the at least three calibration boards are identical; and
the first pattern comprises first circular points or first ellipsoidal points distributed on the calibration board, and a difference between grayscale values of the first circular points or the first ellipsoidal points and a background of the calibration board is greater than the first threshold.

4. The method according to claim 3 , characterized in that:
the first circular points or the first ellipsoidal points are evenly distributed in rows and columns on the calibration board.

5. The method according to claim 3 , characterized in that:
the coordinates of the image formed on the infrared image by the first pattern comprise: coordinates of centers of the first circular points or the first ellipsoidal points in the infrared image; and
the distance between the image formed on the common two-dimensional image by the first pattern and the projection of the three-dimensional cloud point on the common two-dimensional image comprises: a distance between a position of a projection formed on the common two-dimensional image by the centers of the first circular points or the first ellipsoidal points in the infrared image and a position of an image formed on the common two-dimensional image by the centers of the first circular points or the first ellipsoidal points.

6. The method according to claim 1 , characterized in that:
the coordinates of the first pattern in the common two-dimensional image correspond to coordinates of the same first pattern in the infrared image.

7. The method according to claim 3 , characterized in that:
the calibration object comprises four equal-sized calibration boards, wherein the four calibration boards are arranged in two rows and two columns.

8. The method according to claim 1 , characterized in that:
the calibration object carries a pattern indicating a placement direction of the calibration object.

9. The method according to claim 8 , characterized that:
the pattern indicating the placement direction of the calibration object comprises one of: a first number of second circular points arranged in one row and a second number of second circular points arranged in another row; and the first number of second circular points arranged in one column, and the second number of second circular points arranged in another column;
wherein, the second circular points are arranged to form a pattern indicating the placement direction of the calibration object.

10. The method according to claim 1 , characterized in that:
the common two-dimensional image is captured by a first camera device, the infrared image is captured by a second camera device, and the first camera device and the second camera device have different fields of view.

11. The method according to claim 10 , characterized in that:
an image captured by one of the first camera device and the second camera device having a greater field of view includes an image of the entire calibration object;
the image of the calibration object occupies a middle part of the common two-dimensional image, and occupies more than a first ratio of the common two-dimensional image; and
the image of the calibration object occupies a middle part of the infrared image, and occupies more than the first ratio of the infrared image.

12. The method according to claim 11 , characterized in that:
the first ratio is 80%.

13. The method according to claim 1 , characterized in that:
the step of acquiring the common two-dimensional image and the infrared image of the calibration object comprises simultaneously capturing one common two-dimensional image and one infrared image of the calibration object.

14. The method according to claim 1 , characterized in that, the step of calculating, according to the parameter obtained from calibration, the coordinates of the image formed on the infrared image by the first pattern, and the depth value of the first pattern, the position of the projection formed on the common two-dimensional image by the three-dimensional point cloud corresponding to the image formed on the infrared image by the first pattern comprises:
according to (1) the coordinates of the image formed on the infrared image by the first pattern, (2) the depth value of the first pattern, and (3) a second camera device's a focal length, principal-point coordinates, and a skew factor obtained from calibration, calculating a position of a point cloud that projects the first pattern to a three-dimensional coordinate system of the second camera device, wherein the common two-dimensional image is captured by a first camera device and the infrared image is captured by the second camera device;
according to a relative rotation matrix and a relative translation matrix between the first camera device and the second camera device that are obtained from calibration and the position of the point cloud of the first pattern in the three-dimensional coordinate system of the second camera device, calculating a position of a point cloud that projects the first pattern to a three-dimensional coordinate system of the first camera device; and
according to a focal length, principal-point coordinates, a skew factor of the first camera device obtained from calibration, and the position of the point cloud of the first pattern in the three-dimensional coordinate system of the first camera device, calculating a position of a point cloud that projects the first pattern to a two-dimensional coordinate system of the common two-dimensional image.

15. The method according to claim 1 , characterized in that, the step of calculating, according to the parameter obtained from calibr

Assignees

• Arcsoft Corp Ltd

Inventors

• Yu Li
• Liu Menghan
• Miao Wang

Classifications

• G06T7/85 Primary

  Stereo camera calibration · CPC title

• G06T7/55

  from multiple images · CPC title

• G06T2207/10024

  Color image · CPC title

• G06T2207/10028

  Range image; Depth image; 3D point clouds · CPC title

• G06T2207/10048

  Infrared image · CPC title