Image analysis – Applications – 3-d or stereo imaging analysis
Reexamination Certificate
1998-10-09
2002-04-23
Mehta, Bhavesh (Department: 2621)
Image analysis
Applications
3-d or stereo imaging analysis
C382S285000, C345S629000
Reexamination Certificate
active
06377701
ABSTRACT:
BACKGROUND OF THE INVENTION
1. Field of the Invention
The present invention relates to a calibration method and device, a device for generating calibration data and a method thereof, and information providing medium; whereby information regarding the position of an object in 3-dimensional space is calculated from 2-dimensional images obtained by imaging devices which perform imaging of the object.
2. Description of the Related Art
There is a known stereo camera system serving as a 3-dimensional position detecting device, in which a plurality (e.g., two) of video cameras (hereafter simply referred to as “cameras”) are used to comprise a so-called stereo camera by which an object in a 3-dimensional space is imaged and the position of the object in the 3-dimensional space is determined based on the 2-dimensional images obtained by the imaging.
Details of a stereo camera system are disclosed in detail in MASATOSHI OKUTOMI and TAKEO KANEIDE: “Stereo Matching Using Multiple Base-Line Lengths”, Institute of Electronics, Information and Communication Engineers Journal D-II, Vol. J75-D-II. No. 8, pp. 1317-1327 (August 1992), and so forth.
With a stereo camera system, an object whose positional information in 3-dimensional space is to be obtained is imaged by multiple cameras, and information regarding the position of the object in 3-dimensional space can be obtained from the positional information of the object projected on a light-receiving plane (hereafter referred to as “screen”) of the photo-electric converting devices (e.g., CCD) of each camera. Accordingly, in the event that there is positional information of an object existing at a certain position within a 3-dimensional space, and an object at that position, the correlated relation with the positional information of the object projected on the screen of each camera (correlated relation of position information) must be determined beforehand. The process of obtaining this correlated relation of position information is referred to as “calibration”, and is carried out by a calibration device.
FIG. 8
 is an external perspective view of a known calibration device for performing calibration. In 
FIG. 8
, pipes 
106
 and 
107
 are included in the same plane in a 3-dimensional space, and a carriage 
108
 is provided so as to smoothly move along the pipes 
106
 and 
107
. Attached to the carriage 
108
 is a stereo camera, comprised of a first camera 
101
 and a second camera 
102
 which have been integrally joined by means of a metal piece 
103
.
The pipes 
106
 and 
107
 are inscribed with scale marks, so as to enable measuring the distance that the carriage 
108
 slides. A plate 
109
 with a square lattice-work pattern drawn thereupon is provided in a direction perpendicular to the direction of sliding of the carriage 
108
. The horizontal direction of the square lattice-work serves as the X-axis, the vertical direction thereof as the Y-axis, and the direction of sliding, i.e., the direction perpendicular to the square lattice-work is the Z-axis. Z>0 holds for the side of the plate 
109
 on which the camera is provided. Such a 3-dimensional coordinates system with the X-axis, Y-axis, and Z-axis, is defined as a “world coordinates” system.
Calibration measurement is performed by shifting the carriage 
108
 carrying the aforementioned stereo camera along the Z-axis, and imaging the plate 
109
 from two positions. 
FIG. 9
 is a diagram describing a case in which imaging is performed from two positions, viewing the device shown in 
FIG. 8
 from directly above.
First, the first camera 
101
 and the second camera 
102
 are fixed at a certain position P
1
, the plate 
109
 is imaged such that the square lattice-work pattern is recorded, following which the first camera 
101
 and the second camera 
102
 are slid along the Z-axis to another position P
2
 by a distance of M by means of sliding the carriage 
108
, and the plate 
109
 is imaged once more. Here, 
FIG. 9
 shows the first camera 
101
 and the second camera 
102
 being moved in a direction away from the plate 
109
, but this direction of sliding may be reversed.
Thus, the 2-dimensional images obtained by means of sliding a stereo camera comprised of the first camera 
101
 and the second camera 
102
 and imaging the plate 
109
 can also be obtained by means of an arrangement wherein the stereo camera is fixed and the plate 
109
 is shifted instead as well, as shown in FIG. 
10
.
That is to say, the same 2-dimensional images can be obtained by an arrangement such as shown in 
FIG. 10
, wherein the first camera 
101
 and the second camera 
102
 are fixed to a certain position P
1
, the plate 
109
 is imaged such that the square lattice-work pattern is recorded, following which the plate 
109
 is slid along the Z-axis by a distance of M to another position P
2
 in a direction away from the first camera 
101
 and the second camera 
102
, and the plate 
109
 is imaged once more at that position.
In 
FIG. 10
, with the lower left corner of the square lattice-work pattern drawn on the plate before moving the plate by a distance of M (first square lattice-work pattern Q
1
) serving as the origin point and also as the origin for the world coordinates system, the position (i, j) on the first square lattice-work pattern Q
1
 for the plate 
109
 is (i, j, O) on the world coordinates. Also, the position (i, j) on the second square lattice-work pattern Q
2
 after the plate 
109
 has been shifted by a distance of M is (i, j, −M) on the world coordinates.
FIG. 11
 shows the first camera 
101
, and the first square lattice-work Q
1
 and second square lattice-work Q
2
 on the plate 
109
. The optical center of the first camera 
101
 is 
01
, and positional information of the object is case upon the screen serving as the light-receiving surface of a CCD 
122
 or the like. For example, let us say that coordinates position (p, q) at the first square lattice-work Q
1
 is projected, and coordinates position (r, s) at the second square lattice-work Q
2
 is projected. Incidentally, the coordinates of positions other than upon the vertical and horizontal lines in the grid can be calculated by interpolation.
Giving the same description once more using world coordinates, 3-dimensional coordinates positions (p, q, O) and (r, s, −M) are projected onto the coordinates position (h, k) on the CCD 
122
. That is, in the event that the 2-dimensional coordinates position (h, k) and the 3-dimensional coordinates positions (p, q, O) and (r, s, −M) are connected by a line N, all points on this line N are projected onto the coordinates position (h, k) on the CCD 
122
.
Accordingly, the line N represents a correlated relation (correlated relation of position information) between positional information of objects in a 3-dimensional space (in this case, coordinates in the world coordinates system), and 2-dimensional information obtained by imaging the object (in this case, coordinates on the 2-dimensional coordinates system on the CCD 
122
).
This line N can be calculated as follows:
(x−r)/(p−r)=(y−s)/(q−s)=(z+M)/M
In the same manner as calculating line N, lines projected on other coordinates positions as collections of points in the 3-dimensional space are also calculated for the other 2-dimensional coordinate systems on the CCD, as well. The same is also carried out with the second camera 
102
.
Thus, by calculating all lines for the first camera 
101
 and the second camera 
102
, calibration of the stereo camera system is completed.
The positional information of an object in the 3-dimensional space can be calculated as follows, using a stereo camera system which has been calibrated as described above.
First, an object is imaged using the stereo camera. For example, let us say that the object 
127
 shown in 
FIG. 127
 is projected at the position (a, b) on the screen 
122
 of the first camera 
101
 and the position (c, d) on the screen 
128
 of the second camera 
102
. The lines 
130
 and 
131
 on the world coordinates system corresponding with the positions (a,
Bayat Ali
Bell Boyd & Lloyd LLC
Mehta Bhavesh
LandOfFree
Calibration method and device, device for generating... does not yet have a rating. At this time, there are no reviews or comments for this patent.
If you have personal experience with Calibration method and device, device for generating..., we encourage you to share that experience with our LandOfFree.com community. Your opinion is very important and Calibration method and device, device for generating... will most certainly appreciate the feedback.
Profile ID: LFUS-PAI-O-2840312