精确的内在相机校准 - 用于 ROS camera_info 消息
Precise Intrinsic Camera Calibration - for ROS camera_info message
我使用 CALDE 工具 (http://www.dlr.de/rm/en/desktopdefault.aspx/tabid-3925/) 进行内部相机校准，它非常强大，我建议大家使用它进行精确的内部相机校准。
Calde 最后给了我一个如下所示的文件：
% CAMERA # 1
%
% Image size:
imagesize_1 = [ 640 480 ]
% Focal length:
fc_1 = [ 537.417 537.311 ]
% Principal point:
cc_1 = [ 314.329 239.206 ]
% Skew (please note Skew = Gamma/ScaleX):
alpha_c_1 = [ 0.00168813 ]
% Distortion (radial, decentering, and thin-prism, if any)
% (only kc_ is present in Bouguets toolbox, in that order)
kc_1 = [ 0.0450068 -0.144093 0.00000 0.00000 0.00000 ]
radial_1 = [ 0.0450068 -0.144093 0.00000 ]
decentering_1 = [ 0.00000 0.00000 0.00000 ]
thinprism_1 = [ 0.00000 0.00000 0.00000 ]
% TCP_T_CAMERA:
TCP_T_CAMERA1 = [ ...
1.00000 0.00000 0.00000 0.00000 ;
0.00000 1.00000 0.00000 0.00000 ;
0.00000 0.00000 1.00000 0.00000 ]
% MAINCAMERA_T_OBJECT:
MAINCAMERA_T_OBJECT(1:3,1:4,1) = [ ...
0.973021 0.118707 -0.197836 -41.0266 ;
-0.0730198 0.971849 0.224004 -11.4107 ;
0.218857 -0.203514 0.954297 662.759 ]
MAINCAMERA_T_OBJECT(1:3,1:4,2) = [ ...
0.999266 0.0369076 0.0102493 76.4807 ;
-0.0381896 0.980636 0.192078 116.143 ;
-0.00296168 -0.192328 0.981326 1084.49 ]
MAINCAMERA_T_OBJECT(1:3,1:4,3) = [ ...
0.992568 -0.0105678 -0.121230 106.536 ;
0.0316905 0.984295 0.173663 -159.832 ;
0.117491 -0.176215 0.977315 1087.52 ]
MAINCAMERA_T_OBJECT(1:3,1:4,4) = [ ...
0.877250 -0.0191499 0.479652 166.351 ;
-0.0501447 0.990082 0.131240 -29.7432 ;
-0.477408 -0.139182 0.867589 988.947 ]
MAINCAMERA_T_OBJECT(1:3,1:4,5) = [ ...
0.527366 -0.00319134 0.849632 144.484 ;
-0.124694 0.988874 0.0811117 -24.5847 ;
-0.840438 -0.148719 0.521101 969.772 ]
MAINCAMERA_T_OBJECT(1:3,1:4,6) = [ ...
0.891724 0.0522552 -0.449552 94.2428 ;
0.0213817 0.987339 0.157179 -0.458213 ;
0.452074 -0.149773 0.879317 1082.62 ]
MAINCAMERA_T_OBJECT(1:3,1:4,7) = [ ...
0.693097 0.0722656 -0.717212 76.8902 ;
0.0699173 0.983531 0.166666 -17.6251 ;
0.717445 -0.165661 0.676631 1050.90 ]
MAINCAMERA_T_OBJECT(1:3,1:4,8) = [ ...
0.985416 0.0898682 -0.144496 -31.7523 ;
-0.0664750 0.984994 0.159272 140.445 ;
0.156641 -0.147344 0.976603 1565.78 ]
MAINCAMERA_T_OBJECT(1:3,1:4,9) = [ ...
0.992972 0.0389569 -0.111755 -211.173 ;
-0.0133414 0.975099 0.221369 228.341 ;
0.117596 -0.218323 0.968765 1974.26 ]
MAINCAMERA_T_OBJECT(1:3,1:4,10) = [ ...
0.997039 0.0730305 -0.0240702 -26.9394 ;
-0.0675975 0.981645 0.178339 -45.1356 ;
0.0366525 -0.176183 0.983675 706.637 ]
MAINCAMERA_T_OBJECT(1:3,1:4,11) = [ ...
0.998929 0.0272116 -0.0374343 -5.70587 ;
-0.0215852 0.989451 0.143250 -2.35603 ;
0.0409375 -0.142289 0.988978 647.193 ]
MAINCAMERA_T_OBJECT(1:3,1:4,12) = [ ...
0.987303 0.121919 -0.101829 -177.281 ;
0.00962968 0.593922 0.804465 91.9028 ;
0.158558 -0.795231 0.585206 1072.55 ]
现在我需要将这些数据存储在如下所示的 ros camera_info 消息中：
看这里：http://docs.ros.org/kinetic/api/sensor_msgs/html/msg/CameraInfo.html
The distortion model used. Supported models are listed in
sensor_msgs/distortion_models.h. For most cameras, "plumb_bob" - a
simple model of radial and tangential distortion - is sufficient.
string distortion_model
The distortion parameters, size depending on the distortion model.
For "plumb_bob", the 5 parameters are: (k1, k2, t1, t2, k3).
float64[] D
Intrinsic camera matrix for the raw (distorted) images.
[fx 0 cx]
K = [ 0 fy cy]
[ 0 0 1]
Projects 3D points in the camera coordinate frame to 2D pixel
coordinates using the focal lengths (fx, fy) and principal point
(cx, cy).
float64[9] K # 3x3 row-major matrix
Rectification matrix (stereo cameras only)
A rotation matrix aligning the camera coordinate system to the ideal
stereo image plane so that epipolar lines in both stereo images are
parallel.
float64[9] R # 3x3 row-major matrix
Projection/camera matrix
[fx' 0 cx' Tx]
P = [ 0 fy' cy' Ty]
[ 0 0 1 0]
By convention, this matrix specifies the intrinsic (camera) matrix
of the processed (rectified) image. That is, the left 3x3 portion
is the normal camera intrinsic matrix for the rectified image.
It projects 3D points in the camera coordinate frame to 2D pixel
coordinates using the focal lengths (fx', fy') and principal point
(cx', cy') - these may differ from the values in K.
For monocular cameras, Tx = Ty = 0. Normally, monocular cameras will
also have R = the identity and P[1:3,1:3] = K.
For a stereo pair, the fourth column [Tx Ty 0]' is related to the
position of the optical center of the second camera in the first
camera's frame. We assume Tz = 0 so both cameras are in the same
stereo image plane. The first camera always has Tx = Ty = 0. For
the right (second) camera of a horizontal stereo pair, Ty = 0 and
Tx = -fx' * B, where B is the baseline between the cameras.
Given a 3D point [X Y Z]', the projection (x, y) of the point onto
the rectified image is given by:
[u v w]' = P * [X Y Z 1]'
x = u / w
y = v / w
This holds for both images of a stereo pair.
float64[12] P # 3x4 row-major matrix
不知道calde的数据如何得到P矩阵？
看来你只有一个摄像头。 P矩阵中：
fx：fc_1 中的第一个值。
fy：fc_1 中的第二个值。
cx 和 cy：cc_1 中的值与 f.
的顺序相同
如果您只有一个摄像头，文档会显示 Tx 和 Ty 为 0。
关于畸变参数，在ROS文档中畸变参数k是径向畸变参数，t是切向畸变参数（铅锤是径向畸变和切向畸变的组合）。由于您的 CALDE 工具仅计算了前两个径向畸变参数，因此您可以按相同的顺序使用它们。
我使用 CALDE 工具 (http://www.dlr.de/rm/en/desktopdefault.aspx/tabid-3925/) 进行内部相机校准，它非常强大，我建议大家使用它进行精确的内部相机校准。
Calde 最后给了我一个如下所示的文件：
% CAMERA # 1
%
% Image size:
imagesize_1 = [ 640 480 ]
% Focal length:
fc_1 = [ 537.417 537.311 ]
% Principal point:
cc_1 = [ 314.329 239.206 ]
% Skew (please note Skew = Gamma/ScaleX):
alpha_c_1 = [ 0.00168813 ]
% Distortion (radial, decentering, and thin-prism, if any)
% (only kc_ is present in Bouguets toolbox, in that order)
kc_1 = [ 0.0450068 -0.144093 0.00000 0.00000 0.00000 ]
radial_1 = [ 0.0450068 -0.144093 0.00000 ]
decentering_1 = [ 0.00000 0.00000 0.00000 ]
thinprism_1 = [ 0.00000 0.00000 0.00000 ]
% TCP_T_CAMERA:
TCP_T_CAMERA1 = [ ...
1.00000 0.00000 0.00000 0.00000 ;
0.00000 1.00000 0.00000 0.00000 ;
0.00000 0.00000 1.00000 0.00000 ]
% MAINCAMERA_T_OBJECT:
MAINCAMERA_T_OBJECT(1:3,1:4,1) = [ ...
0.973021 0.118707 -0.197836 -41.0266 ;
-0.0730198 0.971849 0.224004 -11.4107 ;
0.218857 -0.203514 0.954297 662.759 ]
MAINCAMERA_T_OBJECT(1:3,1:4,2) = [ ...
0.999266 0.0369076 0.0102493 76.4807 ;
-0.0381896 0.980636 0.192078 116.143 ;
-0.00296168 -0.192328 0.981326 1084.49 ]
MAINCAMERA_T_OBJECT(1:3,1:4,3) = [ ...
0.992568 -0.0105678 -0.121230 106.536 ;
0.0316905 0.984295 0.173663 -159.832 ;
0.117491 -0.176215 0.977315 1087.52 ]
MAINCAMERA_T_OBJECT(1:3,1:4,4) = [ ...
0.877250 -0.0191499 0.479652 166.351 ;
-0.0501447 0.990082 0.131240 -29.7432 ;
-0.477408 -0.139182 0.867589 988.947 ]
MAINCAMERA_T_OBJECT(1:3,1:4,5) = [ ...
0.527366 -0.00319134 0.849632 144.484 ;
-0.124694 0.988874 0.0811117 -24.5847 ;
-0.840438 -0.148719 0.521101 969.772 ]
MAINCAMERA_T_OBJECT(1:3,1:4,6) = [ ...
0.891724 0.0522552 -0.449552 94.2428 ;
0.0213817 0.987339 0.157179 -0.458213 ;
0.452074 -0.149773 0.879317 1082.62 ]
MAINCAMERA_T_OBJECT(1:3,1:4,7) = [ ...
0.693097 0.0722656 -0.717212 76.8902 ;
0.0699173 0.983531 0.166666 -17.6251 ;
0.717445 -0.165661 0.676631 1050.90 ]
MAINCAMERA_T_OBJECT(1:3,1:4,8) = [ ...
0.985416 0.0898682 -0.144496 -31.7523 ;
-0.0664750 0.984994 0.159272 140.445 ;
0.156641 -0.147344 0.976603 1565.78 ]
MAINCAMERA_T_OBJECT(1:3,1:4,9) = [ ...
0.992972 0.0389569 -0.111755 -211.173 ;
-0.0133414 0.975099 0.221369 228.341 ;
0.117596 -0.218323 0.968765 1974.26 ]
MAINCAMERA_T_OBJECT(1:3,1:4,10) = [ ...
0.997039 0.0730305 -0.0240702 -26.9394 ;
-0.0675975 0.981645 0.178339 -45.1356 ;
0.0366525 -0.176183 0.983675 706.637 ]
MAINCAMERA_T_OBJECT(1:3,1:4,11) = [ ...
0.998929 0.0272116 -0.0374343 -5.70587 ;
-0.0215852 0.989451 0.143250 -2.35603 ;
0.0409375 -0.142289 0.988978 647.193 ]
MAINCAMERA_T_OBJECT(1:3,1:4,12) = [ ...
0.987303 0.121919 -0.101829 -177.281 ;
0.00962968 0.593922 0.804465 91.9028 ;
0.158558 -0.795231 0.585206 1072.55 ]
现在我需要将这些数据存储在如下所示的 ros camera_info 消息中： 看这里：http://docs.ros.org/kinetic/api/sensor_msgs/html/msg/CameraInfo.html
The distortion model used. Supported models are listed in
sensor_msgs/distortion_models.h. For most cameras, "plumb_bob" - a
simple model of radial and tangential distortion - is sufficient.
string distortion_model
The distortion parameters, size depending on the distortion model.
For "plumb_bob", the 5 parameters are: (k1, k2, t1, t2, k3).
float64[] D
Intrinsic camera matrix for the raw (distorted) images.
[fx 0 cx]
K = [ 0 fy cy]
[ 0 0 1]
Projects 3D points in the camera coordinate frame to 2D pixel
coordinates using the focal lengths (fx, fy) and principal point
(cx, cy).
float64[9] K # 3x3 row-major matrix
Rectification matrix (stereo cameras only)
A rotation matrix aligning the camera coordinate system to the ideal
stereo image plane so that epipolar lines in both stereo images are
parallel.
float64[9] R # 3x3 row-major matrix
Projection/camera matrix
[fx' 0 cx' Tx]
P = [ 0 fy' cy' Ty]
[ 0 0 1 0]
By convention, this matrix specifies the intrinsic (camera) matrix
of the processed (rectified) image. That is, the left 3x3 portion
is the normal camera intrinsic matrix for the rectified image.
It projects 3D points in the camera coordinate frame to 2D pixel
coordinates using the focal lengths (fx', fy') and principal point
(cx', cy') - these may differ from the values in K.
For monocular cameras, Tx = Ty = 0. Normally, monocular cameras will
also have R = the identity and P[1:3,1:3] = K.
For a stereo pair, the fourth column [Tx Ty 0]' is related to the
position of the optical center of the second camera in the first
camera's frame. We assume Tz = 0 so both cameras are in the same
stereo image plane. The first camera always has Tx = Ty = 0. For
the right (second) camera of a horizontal stereo pair, Ty = 0 and
Tx = -fx' * B, where B is the baseline between the cameras.
Given a 3D point [X Y Z]', the projection (x, y) of the point onto
the rectified image is given by:
[u v w]' = P * [X Y Z 1]'
x = u / w
y = v / w
This holds for both images of a stereo pair.
float64[12] P # 3x4 row-major matrix
不知道calde的数据如何得到P矩阵？
看来你只有一个摄像头。 P矩阵中：
fx：fc_1 中的第一个值。
fy：fc_1 中的第二个值。
cx 和 cy：cc_1 中的值与 f.
的顺序相同如果您只有一个摄像头，文档会显示 Tx 和 Ty 为 0。
关于畸变参数，在ROS文档中畸变参数k是径向畸变参数，t是切向畸变参数（铅锤是径向畸变和切向畸变的组合）。由于您的 CALDE 工具仅计算了前两个径向畸变参数，因此您可以按相同的顺序使用它们。