Automatic scene calibration
Automatic scene calibration
 CN 103,718,213 A
 Filed: 01/14/2013
 Published: 04/09/2014
 Est. Priority Date: 01/13/2012
 Status: Active Application
First Claim
1. under threedimensional environment, calibrate a method for threedimensional flight time imaging system, described method comprises the steps:
 A) determine the reference quadrature virtual threedimensional coordinate system (M of 3D imaging system _{c}), described have level, vertical and degree of depth axle with reference to quadrature virtual threedimensional coordinate system (Mc), wherein said transverse axis and described Zaxis align with the transverse axis and the Zaxis that are positioned at the sensor of described 3D imaging system respectively, and the transverse axis by described sensor of described degree of depth axle and described sensor and the defined planar quadrature of Zaxis;
B) at virtual coordinate system (M _{c}) the middle vertical direction (V that obtains real world _{w});
C) with respect to described reference frame, determine real world threedimensional orthogonal coordinate system (M _{w}), described real world threedimensional orthogonal coordinate system (M _{w}) there is level, vertical and degree of depth axle, wherein rotate the described Zaxis described vertical direction (V that makes it to align _{w});
D) determine in scene a bit as described real world threedimensional orthogonal coordinate system (M _{w}) new initial point;
E) from described virtual threedimensional coordinate system (M _{c}) initial point to the point that is defined as the new initial point of described scene, derive translation vector (T _{v});
F) derive rotation matrix (M _{r}), for being described real world threedimensional system of coordinate by described virtual threedimensional coordinate system transformation, and;
G) derive the calibration matrix (M of described 3D imaging system _{c2w}) as the rotation matrix by the translation of described translation vector institute;
It is characterized in that, described method also comprises the Zaxis by described real world threedimensional system of coordinate and the defined plane of degree of depth axle and described virtual threedimensional coordinate system (M _{c}thereby) virtual Zaxis and the coplanar step of the defined planar registration of virtual depth axle.
Chinese PRB Reexamination
Abstract
Described herein is a method of calibrating a threedimensional imaging system (300). During calibration, a position (340) and an orientation (330) of the threedimensional imaging system is determined with respect to a first parameter comprising a real world vertical direction (Vw) and to a second parameter comprising an origin of a threedimensional scene captured by the imaging system. The first and second parameters are used to derive a calibration matrix (MC2w) which is used to convert measurements (360) from a virtual coordinate system (Mc) of the threedimensional imaging system into a real coordinate system (Mw) related to the real world. The calibration matrix (MC2w) is used to rectify measurements (360) prior to signal processing (380). An inverse calibration matrix (Mw2c) is also determined. Continuous monitoring (370) and adjustment of the setup of the threedimensional imaging system is carried out and the calibration matrix (Mc2w) and its inverse (Mw2c) are adjusted accordingly.

12 Citations
Use the method and computer program product of flat mirror calibration stereo imaging system  
Patent #
CN 109,155,070 A
Filed 03/07/2017

Current Assignee

Method and system for compensating for vehicle calibration errors  
Patent #
CN 111,532,257 A
Filed 04/21/2020

Current Assignee

Method and device for determining external parameter of stereoscopic camera  
Patent #
US 11,010,924 B2
Filed 09/30/2017

Current Assignee
N/A

Pavement detection method and device based on multiline laser radar  
Patent #
CN 110,109,146 A
Filed 04/30/2019

Current Assignee

A kind of AR sand table calibrating patterns computational methods for capableing of automatic deviation correction  
Patent #
CN 108,320,311 A
Filed 01/19/2018

Current Assignee

A kind of definite method and device of stereo camera external parameter  
Patent #
CN 107,945,234 A
Filed 10/12/2016

Current Assignee

Based on threedimensional environment imaging method and device  
Patent #
CN 106,846,410 A
Filed 12/20/2016

Current Assignee

Antidote, device and the virtual reality device of camera parameter  
Patent #
CN 108,282,651 A
Filed 12/18/2017

Current Assignee

Video camera reference method only using plane reference object image  
Patent #
CN 1,801,953 A
Filed 12/15/2005

Current Assignee

METHODS AND SYSTEMS FOR DETERMINING THE POSE OF A CAMERA WITH RESPECT TO AT LEAST ONE OBJECT OF A REAL ENVIRONMENT  
Patent #
US 20100208057A1
Filed 02/13/2009

Current Assignee
N/A

3D TIMEOFFLIGHT CAMERA SYSTEM AND POSITION/ORIENTATION CALIBRATION METHOD THEREFOR  
Patent #
US 20110205340A1
Filed 04/27/2009

Current Assignee
N/A

Self calibration of extrinsic camera parameters for a vehicle camera  
Patent #
CN 102,037,735 A
Filed 05/08/2009

Current Assignee

16 Claims

1. under threedimensional environment, calibrate a method for threedimensional flight time imaging system, described method comprises the steps:

A) determine the reference quadrature virtual threedimensional coordinate system (M of 3D imaging system _{c}), described have level, vertical and degree of depth axle with reference to quadrature virtual threedimensional coordinate system (Mc), wherein said transverse axis and described Zaxis align with the transverse axis and the Zaxis that are positioned at the sensor of described 3D imaging system respectively, and the transverse axis by described sensor of described degree of depth axle and described sensor and the defined planar quadrature of Zaxis;
B) at virtual coordinate system (M _{c}) the middle vertical direction (V that obtains real world _{w});
C) with respect to described reference frame, determine real world threedimensional orthogonal coordinate system (M _{w}), described real world threedimensional orthogonal coordinate system (M _{w}) there is level, vertical and degree of depth axle, wherein rotate the described Zaxis described vertical direction (V that makes it to align _{w});
D) determine in scene a bit as described real world threedimensional orthogonal coordinate system (M _{w}) new initial point;
E) from described virtual threedimensional coordinate system (M _{c}) initial point to the point that is defined as the new initial point of described scene, derive translation vector (T _{v});
F) derive rotation matrix (M _{r}), for being described real world threedimensional system of coordinate by described virtual threedimensional coordinate system transformation, and;
G) derive the calibration matrix (M of described 3D imaging system _{c2w}) as the rotation matrix by the translation of described translation vector institute;
It is characterized in that, described method also comprises the Zaxis by described real world threedimensional system of coordinate and the defined plane of degree of depth axle and described virtual threedimensional coordinate system (M _{c}thereby) virtual Zaxis and the coplanar step of the defined planar registration of virtual depth axle.


2. method according to claim 1, is characterized in that, step g) also comprises from described calibration matrix (M _{c2w}) derive against calibration matrix (M _{w2c}), for described real world threedimensional system of coordinate is transformed to virtual threedimensional coordinate system.

3. method according to claim 1 and 2, is characterized in that, step b) comprises will derive the vertical direction (V of described real world _{w}) as the opposite vector that uses the determined gravity vector of measurement mechanism, described measurement mechanism comprises at least one Inertial Measurement Unit.

4. method according to claim 1 and 2, is characterized in that, step b) comprises that the normal of one plane from scene derives the vertical direction (V of described real world _{w}), the plane in described scene is determined by following steps:

I) use the described 3D imaging system in the first orientation to catch scene; II) with plane fitting algorithm, determine a plurality of planes in described scene;
AndIII) determine that reference planes in described scene are as ground.


5. method according to claim 4, is characterized in that, step III) comprise that will be best meeting following one is defined as described reference planes:
 statistical model;
Maximized surface;
And minimal surface is longpending.
 statistical model;

6. method according to claim 4, is characterized in that, step III) comprise the combination as following, determine described reference planes:
 the statistical model of the principal component analysis (PCA) of described scene;
Maximized surface;
And minimal surface is longpending.
 the statistical model of the principal component analysis (PCA) of described scene;

7. method according to claim 1 and 2, is characterized in that, step b) comprises that from described scene specific user'"'"'s stance derives the vertical direction (V of described real world _{w}), described vertical direction (V _{w}) be the direction of aliging with the user'"'"'s who stands with predetermined calibration attitude described vertical direction.

8. method according to claim 7, is characterized in that, also comprise the steps:
 to come leadout level axle and Zaxis from described predetermined calibration attitude, and by described real world threedimensional system of coordinate (M _{w}) align with derived level and Zaxis.

9. according to the method described in any one in claims 1 to 3, it is characterized in that, step b) comprises derives the vertical direction of described real world (V in the edge from detecting described scene _{w}).

10. method according to claim 1 and 2, is characterized in that, step b) comprises by combining two or more following steps, comes vertical direction (V _{w}) be optimized:
 the vertical direction (V that derives described real world _{w}) as the opposite vector that uses the determined gravity vector of measurement mechanism;
From described scene, the normal of one plane is derived the vertical direction (V of described real world _{w});
Specific user'"'"'s stance from described scene derives the vertical direction (V of described real world _{w}), described vertical direction (V _{w}) be the direction of aliging with the user'"'"'s that stands with predetermined calibration attitude vertical direction;
And from scene detection to edge obtain real world vertical direction (V _{w})
 the vertical direction (V that derives described real world _{w}) as the opposite vector that uses the determined gravity vector of measurement mechanism;

11. according to the method described in any one in the claims, it is characterized in that, step d) comprises while determining new initial point and adopts one of following methods:
 a predetermined point in usage space;
The extreme lower position of definition user'"'"'s each point;
Be positioned at a bit in the plane detecting;
And the position of predetermined object in described scene.
 a predetermined point in usage space;

12. according to the method described in any one in the claims, it is characterized in that, also comprises when detecting in following and when at least one item changes, automatically improves calibration matrix (M _{c2w}) step:
 with respect to the position of the 3D imaging system of threedimensional environment and orientation, virtual coordinate system (M _{c}) in by (the M of described realworld coordinates system _{w}) at least one axles of two defined planar registration of axle.

13. methods according to claim 12, is characterized in that, also comprise by controlling the maneuvering system of bearing threedimensional imaging system, by described virtual coordinate system (M _{c}) at least Zaxis and the (M of described realworld coordinates system _{w}) the step of described YZ planar registration.

14. 1 kinds of degree of depth sensing imaging systems with Inertial Measurement Unit, described degree of depth sensing imaging system is according to operating according to the method described in abovementioned arbitrary claim.

15. 1 kinds of degree of depth sensing imaging systems with powerdriven tool, following at least one for adjusting:
 according to position and the orientation of the degree of depth sensing imaging system of the method operation described in any according to claim 1 to 13.

16. 1 kinds of threedimensional flight time imaging systems, have according to the truing tool of method operation described in any in claim 1 to 13.
Specification(s)