Camera calibration application

• US 9,870,704 B2
• Filed: 06/20/2012
• Issued: 01/16/2018
• Est. Priority Date: 06/20/2012
• Status: Active Grant

• Current Assignee: Conduent Business Services, LLC (Conduent, Inc.)
• Original Assignee: Conduent Business Services, LLC (Conduent, Inc.)
• Inventors: Hoover, Martin Edward, Jackson, David Martin Todd, Wu, Wencheng, Kozitsky, Vladimir
• Primary Examiner(s): Rao, Andy
• Assistant Examiner(s): Fereja, Samuel D
• Application Number:
  US13/527,673
• Publication Number:
  US 20130342706A1
• Time in Patent Office: 2,036 Days
• Field of Search: None
• CPC Class Codes: G06T 2207/10016 Video; Image sequence
• CPC Class Codes: G06T 2207/30204 Marker
• CPC Class Codes: G06T 2207/30236 Traffic on road, railway or crossing
• CPC Class Codes: G06T 7/80 Analysis of captured images to determine intrinsic or extrinsic camera parameters, i.e. camera calibration
• CPC Class Codes: G08G 1/0175 by photographing vehicles, e.g. when violating traffic rules
• CPC Class Codes: H04N 17/002 for television cameras

Abstract

A method and system for camera calibration comprises configuring a calibration target comprising calibration reflectors on a test vehicle. Video of a test scene is collected. Next the test vehicle is identified as it enters the test scene and recorded as it passes through the test scene. The position of the calibration target in each frame of the video is determined and the corresponding individual position of each calibration reflector for each frame of the recorded frames is used to construct a camera calibration map to calibrate the video camera.

Claims

1. A method for camera calibration comprising:

configuring a calibration target comprising a rigid platform and a plurality of internally lit calibration reflectors said plurality of calibration reflectors configured in a grid of at least four of said calibration reflectors mounted on said rigid platform in a plane parallel to a road and mounted to a rear of a test vehicle wherein said calibration target is towed by said test vehicle;

recording a plurality of video frames of said calibration target comprising said plurality of calibration reflectors as they pass through a test scene wherein a position of said calibration target and a position of each of said calibration reflectors in each frame of said plurality of recorded frames is collected;

identifying a centroid for each of said calibration reflectors in each frame of said recorded video;

identifying a region within said video frame that encloses all of said calibration reflectors for each frame of said collected video; and

constructing a camera calibration map to calibrate said video camera wherein constructing said camera calibration map further comprises;

selecting an arbitrary fixed reference point in said recorded video frames as an origin in a real world coordinate system;

setting a coordinate of said origin, in said real world coordinate system, associated with a plane parallel to said road, to zero;

constructing a first calibration map for each frame of said recorded video according to a relative position of said centroids for each of said calibration reflectors according to said selected origin; and

averaging all of said first calibration maps.

2. The method of claim 1 further comprising:

collecting said video of said test scene with a video camera; and

applying an automated license plate recognition check for each vehicle that enters said test scene until said test vehicle is identified.

3. The method of claim 1 wherein said calibration target is configured to further comprise at least two risers connected to said rigid platform with associated calibration reflectors thereon in a plane perpendicular to the road, wherein a relative location of each of said plurality of calibration reflectors within said grid is known, and said grid covers as much of a field of view of said test scene as possible.

4. The method of claim 1 wherein constructing a camera calibration map further comprises:

constructing an optimal mapping with a required offset such that said optimal mapping satisfies a least squares approximation given by said selected origin and said centroid of each or said calibration reflectors;

applying said optimal mapping to estimate, real world, coordinates of a next frame of collected video;

computing a corresponding average offset from known relative positions of said centroids;

iterating said process for each of said collected video frames;

removing data with a large average offset according to a rigid body constraint; and

constructing a final calibration map according to said remaining data.

5. The method of claim 1 wherein said collected video comprises a traffic surveillance video feed.

6. The method of claim 5 wherein said test scene comprises one of:

an open roadway;

or

an open traffic intersection.

7. A system for camera calibration comprising:

a calibration target comprising a rigid platform and a plurality of internally lit calibration reflectors said plurality of calibration reflectors configured in a grid of at least four calibration reflectors mounted on said rigid platform in a plane parallel to a road and mounted to the rear of a test vehicle wherein said calibration target is towed by said test vehicle;

a video camera configured to collect video of a test scene;

a test vehicle identification module configured to identify said test vehicle as it enters said test scene and record a plurality of video frames of said calibration target comprising said plurality of calibration reflectors as they pass through said test scene;

a calibration point identification module configured to detect the position of said calibration target in each frame of said plurality of recorded frames, identify a corresponding individual position of each of said calibration reflectors for each frame of said plurality of recorded frames, identify a centroid for each of said calibration reflectors in each frame of said plurality of recorded frames and a region within said video frame that encloses all of said calibration reflectors for each of said plurality of recorded frames; and

a camera calibration construction module configured select an arbitrary fixed reference point in said collected video as an origin in a real world coordinate system, set a coordinate of said origin, in a real world coordinate system, associated with a plane parallel to said road to zero construct a first calibration map for each of said plurality of recorded frames according to a relative position of said centroids for each of said calibration reflectors according to said selected origin, and average all of said first calibration maps to construct a camera calibration map.

8. The system of claim 7 wherein said test vehicle identification module if further configured to apply an automated license plate recognition check for each vehicle that enters said test scene until said test vehicle is identified.

9. The system of claim 7 wherein said calibration target is configured to further comprise at least two risers connected to said rigid platform with associated calibration reflectors thereon in a plane perpendicular to the road, wherein a relative location of each of said plurality of calibration reflectors within said grid is known and said grid covers as much of a field of view of said test scene as possible.

10. The system of claim 9 wherein said camera calibration construction module is further configured for:

constructing an optimal mapping with a required offset such that said optimal mapping satisfies a least squares approximation given by said selected origin and said centroid of each or said calibration reflectors;

applying said optimal mapping to estimate real world coordinates of a next frame of collected video;

computing a corresponding average offset from known relative positions of said centroids;

iterating said process for each of said collected video frames;

removing data with a large average offset according to a rigid body constraint; and

constructing a final calibration map according to said remaining data.

11. The system of claim 7 wherein said collected video comprises a traffic surveillance video feed.

12. The system of claim 11 wherein said test scene comprises one of:

an open roadway;

or

an open traffic intersection.

13. A method for camera calibration comprising:

configuring a calibration target comprising a rigid platform and a plurality of internally lit calibration reflectors said plurality of calibration reflectors configured in a grid of at least four calibration reflectors mounted on said rigid platform in a plane parallel to a road and mounted to the rear of a test vehicle wherein said calibration target is towed by said test vehicle and wherein said calibration target is configured to be a grid of calibration reflectors with at least two riser with associated calibration reflectors thereon wherein a relative location of each of said plurality of calibration reflectors is known;

collecting video of a test scene with a video camera;

identifying said test vehicle as it enters said test scene by applying an automated license plate recognition check for each vehicle that enters said test scene until said test vehicle is identified;

recording a plurality of video frames of said calibration target comprising said plurality of calibration reflectors as they pass through said test scene;

detecting the position of said calibration target in each frame of said plurality of recorded frames;

identifying a corresponding individual position of a centroid of each of said calibration reflectors for each frame of said plurality of recorded frames;

identifying a region that encloses all of said calibration reflectors in each frame of said collected video; and

constructing a camera calibration map to calibrate said video camera wherein constructing said camera calibration map further comprises;

selecting an arbitrary fixed reference point in said collected video as an origin in a real world coordinate system;

setting a coordinate of said origin, in a real world coordinate system, associated with a plane parallel to said road to zero;

constructing a first calibration map for each frame of said collected video according to a relative position of said centroids for each of said calibration reflectors according to said selected origin; and

averaging all of said first calibration maps.

14. The method of claim 13 wherein constructing a camera calibration map further comprises:

constructing an optimal mapping such that said optimal mapping satisfies a least squares approximation given by said selected origin and said centroid of each or said calibration reflectors;

applying said optimal mapping to estimate real world coordinates of a next frame of collected video;

computing a corresponding average offset from known relative positions of said centroids;

iterating said process for each of said collected video frames;

removing data with a large average offset according to a rigid body constraint; and

constructing a final calibration map according to said remaining data.

15. The method of claim 13 wherein said test scene comprises one of:

an open roadway;

or

an open traffic intersection.