Computer vision system for position monitoring in three dimensions using non-coplanar light sources attached to a monitored object

• US 5,227,985 A
• Filed: 08/19/1991
• Issued: 07/13/1993
• Est. Priority Date: 08/19/1991
• Status: Expired due to Term

• Current Assignee: University of Maryland A Non-Profit Organization of MD
• Original Assignee: University of Maryland
• Inventors: DeMenthon, Daniel F.
• Primary Examiner(s): Harvey, Jack B.
• Assistant Examiner(s): BUROKER, BRIAN
• Application Number:
  US07/747,124
• Time in Patent Office: 694 Days
• Field of Search: 364/559, 364/516, 364/410, 364/413.01, 364/413.13, 340/706, 358/105, 358/107, 395/127, 395/152
• US Class Current: 702/153
• CPC Class Codes: G01S 5/163 Determination of attitude using inertial means G01C9/00; control of attitude G05D1/49

Abstract

A sensing system for monitoring the position and orientation of a rigid object (20). At least 4 point light sources (24) are mounted on the surface of the object (20) in a noncoplanar arrangement. A single electronic camera (26) captures images (59) of the point light sources (24). Locations of the images (59) of the light sources (24) are detected in each video image, and a computer runs a task using these locations to obtain close approximations of the rotation matrix and translation vector (33) of the object (20) in a camera coordinate system (74) at video rate. The object is held by an operator (90) for three-dimensional cursor (94) control and interaction with virtual reality scenes (96) on computer displays (88), and for remote interactive control of teleoperated mechanisms.

Claims

1. In an apparatus for monitoring the position and orientation of a rigid three-dimensional (3-D) object in space comprising:

a single electronic camera having an image plane, a center of projection, an optical axis, and a camera reference coordinate frame being centered at the center of projection with x and y axes parallel to the image plane and a z-axis parallel to the optical axis,

the single electronic camera producing an analog video signal;

a plurality of at least four light sources rigidly attached to the rigid 3-D object in a noncoplanar arrangement, these light sources projecting bright spots onto the image plane of the camera;

processing means for processing the analog video signal and determining a list of bright spot x-coordinates and y-coordinates in the camera reference coordinate frame;

computing means including memory means, output means and display means;

the computing means including pose computing means for approximately computing a translation vector and a rotation matrix of the rigid 3-D object in the camera reference coordinate frame;

the output means modifying images on the display means;

the improvement comprising;

the arrangement of light sources being any noncoplanar arrangement so long as the light sources are not all located along three orthogonal lines.

2. The improvement of the apparatus of claim 1, further comprising coupling means between the rigid 3-D object and a body part of an operator, whereby motions of the body part of the operator are transmitted to the rigid 3-D object and are monitored by the single electronic camera.

3. The improvement of the apparatus of claim 2, wherein the coupling means is a means for hand operation.

4. The improvement of the apparatus of claim 2, wherein the coupling means is a means for head operation.

5. The improvement of the apparatus of claim 2, further comprising structure coordinate lists stored in the memory means for describing a geometry of 3-D structures;

the computing means further including perspective projection means repeatedly combining the translation vector and the rotation matrix with the structure coordinate lists to output a perspective image of the 3-D structures onto the display means.

6. The improvement of the apparatus of claim 5, wherein:

one of the 3-D structures is a 3-D cursor, whereby a

perspective image of the 3-D cursor is displayed on the display means and interactively emulates translations and rotations of the body part of the operator.

7. The improvement of the apparatus of claim 5, wherein:

one of the 3-D structures is a 3-D scene, whereby a perspective image of the 3-D scene is displayed on the display means in front of the eyes of the operator from a point of view which can be interactively modified by the motions of the body part of the operator.

8. The improvement of the apparatus of claim 1, wherein the light sources are light emitting diodes.

9. The improvement of the apparatus of claim 1, wherein the light sources are secondary light sources composed of small flat reflective elements which reflect light from a single primary light source.

10. The improvement of the apparatus of claims 1, wherein:

the light sources differ in sizes;

the computing means further includes means for measuring the sizes of the bright spots;

the computing means further includes means for matching large light sources to large bright spots and small light sources to small bright spots.

11. The improvement of the apparatus of claim 1, wherein:

the light sources are positioned along lines and form known distance ratios with respect to each other;

the computing means further includes means for detecting alignments among the bright spots, means for measuring distance ratios between the bright spots along these alignments; and

means for matching the light sources to the bright spots according to their distance ratios.

12. The improvement of the apparatus of claim 1, wherein:

an object reference coordinate frame for the rigid 3-D object is centered at one reference light source, this reference light source projecting onto the image plane at a reference bright spot m having an x-coordinate and a y-coordinate in the camera reference coordinate frame, the locations of light sources other than the reference light source in the rigid 3-D object being defined by secondary light source coordinates in the object reference coordinate frame;

the memory means stores a precomputed pseudo-inverse matrix obtained by pseudo-inversion of a matrix in which three elements of each row are the secondary light source coordinates;

the computing means further includes means for;

computing a vector of x-values and a vector of y-values by subtracting the x-coordinate and y-coordinate of the reference bright spot m from the list of bright spot x-coordinates and y-coordinates respectively in the camera reference coordinate frame;

multiplying the pseudo-inverse matrix by the vector of x-values to provide a vector V_x and multiplying the precomputed pseudo-inverse matrix by the vector of y-values to provide a vector V_y,

constructing the rotation matrix of the rigid 3-D object in the single electronic camera frame of reference by taking as first row the vector V_x after normalization, as second row the vector V_y after normalization, and as third row the cross product of the vectors of the first two rows,

finding the translation vector of the rigid 3-D object as the division of vector Om from the center of projection O of the single electronic camera to the reference bright spot m by the average of the approximately equal lengths of the vector V_x and the vector V_y.

13. The improvement of the apparatus of claim 12, wherein the pose computing means further includes prediction means for finding predicted bright spot locations on the image plane by extrapolations of data obtained at previous instants and uses the predicted bright spot locations in replacement to actual bright spot locations every time the actual bright spot locations fail to be detected from the analog video signal of the single electronic camera.

14. An apparatus for three-dimensional (3-D) cursor control by an operator comprising:

a single electronic camera having an image plane, a center of projection, an optical axis, and a camera reference coordinate frame being centered at the center of projection with x and y axes parallel to the image plane and a z-axis parallel to the optical axis, the single electronic camera producing an analog video signal;

a plurality of at least four light sources configured in any noncoplanar arrangement so long as the light sources are not all located along three orthogonal lines, the light sources having a light source reference coordinate frame and light source coordinates in this object reference coordinate frame, these light sources projecting bright spots onto the image plane of the camera;

means for processing the analog video signal and determining a list of bright spot locations in the camera reference coordinate frame;

handle means for allowing an operator to hold and move the plurality of light sources freely in space;

computing means for;

repeatedly combining the bright spot locations with the light source coordinates;

repeatedly producing a rotation matrix and a translation vector in the camera reference coordinate frame;

repeatedly outputting onto display means in front of the operator'"'"'s eyes a perspective image of a 3-D virtual cursor located in space at a position and orientation defined by the rotation matrix and the translation vector.

15. A sensing system for producing at successive time instants digital signals expressing positions and orientations of a rigid three-dimensional (3-D) object in space comprising:

a single electronic camera having an image plane, a center of projection, an optical axis, and a camera reference coordinate frame being centered at the center of projection with x and y axes parallel to the image plane and a z-axis parallel to the optical axis, the single electronic camera producing an analog video signal;

at least four noncoplanar light sources rigidly attached to the rigid 3D object in any noncoplanar arrangement so long as the light sources are not all located along three orthogonal lines, the light sources having light source coordinates in an object reference coordinate frame of the rigid 3D object, the light sources projecting bright spots onto the image plane of the camera;

means for processing the analog video signal and determining a list of bright spot locations in the camera reference coordinate frame;

computing means which include memory means and output means;

the memory means storing a precomputed pseudo-inverse matrix depending only on the light source coordinates in the object reference coordinate frame;

the computing means including pose computing means for;

computing a translation vector and a rotation matrix of the rigid 3-D object in the camera reference coordinate frame,

combining the precomputed pseudo-inverse matrix with the list of bright spot locations, by matrix multiplications, vector normalizations and vector cross-products,

and providing to the output means for each frame of the video signal the three coordinates of the translation vector of the rigid 3-D object and the nine elements of the rotation matrix of the rigid 3-D object in digital form.