find_marks_and_pose🔗
Short description🔗
find_marks_and_pose — Extract rectangularly arranged 2D calibration marks from the image and
calculate initial values for the external camera parameters.
Signature🔗
find_marks_and_pose( image Image, region CalPlateRegion, filename.read CalPlateDescr, campar StartCamParam, number StartThresh, number DeltaThresh, number MinThresh, real Alpha, real MinContLength, real MaxDiamMarks, out real RCoord, out real CCoord, out pose StartPose )
Description🔗
find_marks_and_pose is used to determine the
input data for a subsequent camera calibration using a calibration plate
with rectangularly arranged marks (see
Calibration or camera_calibration):
First, the 2D center points
[RCoord,CCoord] of the calibration marks within
the region CalPlateRegion of the input image Image
are extracted and ordered. Secondly, a rough estimate for
the external camera parameters (StartPose) is computed,
i.e., the 3D pose (= position and orientation) of the calibration plate
relative to the camera coordinate system (see create_pose for more
information about 3D poses).
In the input image Image an edge detector is applied
(see edges_image, mode ‘lanser2’) to the region
CalPlateRegion,
which can be found by applying the operator find_caltab.
The filter parameter for this edge detection can be tuned via
Alpha. Use a smaller value for Alpha to achieve
a stronger smoothing effect.
In the edge image closed contours are searched for: The number of
closed contours must correspond to the number of calibration marks
as described in the calibration plate description file
CalPlateDescr and the contours have to be elliptically shaped.
Contours shorter than MinContLength are discarded,
just as contours enclosing regions with a diameter larger than
MaxDiamMarks (e.g., the border of the calibration plate).
For the detection of contours a threshold operator is applied on the resulting amplitudes of the edge detector. All points with a high amplitude (i.e., borders of marks) are selected.
First, the threshold value is set to
StartThresh. If the search for the closed contours or the
successive pose estimate fails, this threshold value is successively
decreased by DeltaThresh down to a minimum value of
MinThresh.
Each of the found contours is refined with subpixel accuracy (see
edges_sub_pix) and
subsequently approximated by an ellipse. The center points of these
ellipses represent a good approximation of the desired 2D image coordinates
[RCoord,CCoord] of the calibration mark center points.
The order of the values within these two tuples must correspond to
the order of the 3D coordinates of the calibration marks in the calibration
plate description file CalPlateDescr, since this fixes the
correspondences between extracted image marks and known model marks
(given by caltab_points)! If a
triangular orientation mark is defined in a corner of the plate by the
plate description file (see gen_caltab), the mark will be detected
and the point order is returned in row-major order beginning with the
corner mark in the (barycentric) negative quadrant with respect to the
defined coordinate system of the plate. Else, if no orientation mark is
defined, the order of the center points is in row-major order beginning at
the upper left corner mark in the image.
Based on the ellipse parameters for each calibration mark, a rough
estimate for the external camera parameters is finally computed.
For this purpose the fixed correspondences between extracted image marks
and known model marks are used. The estimate StartPose
describes the pose of the calibration plate in the camera coordinate
system as required by the operator camera_calibration.
Execution information🔗
Execution information
-
Multithreading type: reentrant (runs in parallel with non-exclusive operators).
-
Multithreading scope: global (may be called from any thread).
-
Processed without parallelization.
Parameters🔗
Image (input_object) singlechannelimage → object (byte / uint2)
Input image.
CalPlateRegion (input_object) region → object
Region of the calibration plate.
CalPlateDescr (input_control) filename.read → (string)
File name of the calibration plate description.
Default: 'caltab_100.descr'
List of values: 'caltab_100mm.descr', 'caltab_10mm.descr', 'caltab_200mm.descr', 'caltab_2500um.descr', 'caltab_30mm.descr', 'caltab_650um.descr', 'caltab_6mm.descr', 'caltab_800mm.descr', 'caltab_big.descr', 'caltab_small.descr'
File extension: .descr
StartCamParam (input_control) campar → (real / integer / string)
Initial values for the internal camera parameters.
StartThresh (input_control) number → (integer)
Initial threshold value for contour detection.
Default: 128
Suggested values: 80, 96, 112, 128, 144, 160
Restriction: StartThresh > 0
DeltaThresh (input_control) number → (integer)
Loop value for successive reduction of
StartThresh.
Default: 10
Suggested values: 6, 8, 10, 12, 14, 16, 18, 20, 22
Restriction: DeltaThresh > 0
MinThresh (input_control) number → (integer)
Minimum threshold for contour detection.
Default: 18
Suggested values: 8, 10, 12, 14, 16, 18, 20, 22
Restriction: MinThresh > 0
Alpha (input_control) real → (real)
Filter parameter for contour detection, see
edges_image.
Default: 0.9
Suggested values: 0.1, 0.2, 0.3, 0.4, 0.5, 0.6, 0.7, 0.8, 0.9, 1.0, 1.1
Value range: 0.2 ≤ Alpha ≤ 2.0
Restriction: Alpha > 0.0
MinContLength (input_control) real → (real)
Minimum length of the contours of the marks.
Default: 15.0
Suggested values: 10.0, 15.0, 20.0, 30.0, 40.0, 100.0
Restriction: MinContLength > 0.0
MaxDiamMarks (input_control) real → (real)
Maximum expected diameter of the marks.
Default: 100.0
Suggested values: 50.0, 100.0, 150.0, 200.0, 300.0
Restriction: MaxDiamMarks > 0.0
RCoord (output_control) real-array → (real)
Tuple with row coordinates of the detected marks.
CCoord (output_control) real-array → (real)
Tuple with column coordinates of the detected marks.
StartPose (output_control) pose → (real / integer)
Estimation for the external camera parameters.
Number of elements: 7
Example🔗
(HDevelop)
* Read calibration image.
read_image(Image, 'calib/calib_distorted_01')
* Find calibration pattern.
find_caltab(Image,CalPlate, 'caltab_100mm.descr', 3, 112, 5)
* Find calibration marks and start pose.
find_marks_and_pose(Image, CalPlate, 'caltab_100mm.descr' , \
['area_scan_division', 0.008, 0.0, \
0.000011, 0.000011, 384, 288, 640, 512], \
128, 10, 18, 0.9, 15.0, 100.0, RCoord, CCoord, StartPose)
HTuple StartCamPar, RCoord, CCoord, StartPose\;
// Read calibration image.
HImage Image("calib/calib_distorted_01")\;
// Find calibration pattern.
HRegion CalPlate = Image.FindCaltab("caltab_100mm.descr", 3, 112, 5)\;
// Read internal camera parameters from file.
read_cam_par("campar.dat", &StartCamPar)\;
// Find calibration marks and start pose.
RCoord = Image.FindMarksAndPose(CalPlate, "caltab_100mm.descr", StartCamPar,
128, 10, 18, 0.9, 15.0, 100.0,
&RCoord, &CCoord, &StartPose)\;
Result🔗
find_marks_and_pose returns 2 (H_MSG_TRUE) if all parameter values are
correct and an estimation for the external camera parameters has been
determined successfully. If necessary, an exception is raised.
Combinations with other operators🔗
Combinations
Possible predecessors
Possible successors
See also
find_caltab, camera_calibration, disp_caltab, sim_caltab, read_cam_par, read_pose, create_pose, pose_to_hom_mat3d, caltab_points, gen_caltab, edges_sub_pix, edges_image
Module🔗
Foundation