# -*- coding: utf-8 -*-
""" Class to do stateful video calibration of a mono camera. """
import copy
import logging
import sksurgeryimage.calibration.point_detector as pd
import sksurgerycalibration.video.video_calibration_driver_base as vdb
import sksurgerycalibration.video.video_calibration_data as cd
import sksurgerycalibration.video.video_calibration_params as cp
import sksurgerycalibration.video.video_calibration_utils as cu
import sksurgerycalibration.video.video_calibration_wrapper as vc
LOGGER = logging.getLogger(__name__)
[docs]class MonoVideoCalibrationDriver(vdb.BaseVideoCalibrationDriver):
""" Class to do stateful video calibration of a mono camera. """
def __init__(self,
point_detector: pd.PointDetector,
minimum_points_per_frame: int
):
"""
Stateful class for mono video calibration.
:param point_detector: Class derived from PointDetector
:param minimum_points_per_frame: Minimum number to accept frame
"""
super().__init__(minimum_points_per_frame)
self.point_detector = point_detector
# Create data holders, and parameter holders, specific to Mono.
calibration_data = cd.MonoVideoData()
calibration_params = cp.MonoCalibrationParams()
# Pass them to base class, so base class can access them.
self._init_internal(calibration_data, calibration_params)
[docs] def grab_data(self,
image,
device_tracking=None,
calibration_object_tracking=None):
"""
Extracts points, by passing it to the PointDetector.
This will throw various exceptions if the input data is invalid,
but will return empty arrays if no points were detected.
So, no points is not an error. Its an expected condition.
:param image: RGB image.
:param device_tracking: transformation for the tracked device
:param calibration_object_tracking: transformation of tracked
calibration object
:return: The number of points grabbed.
"""
number_of_points = 0
ids, object_points, image_points = \
self.point_detector.get_points(image)
if ids is not None and ids.shape[0] >= self.minimum_points_per_frame:
ids, image_points, object_points = \
cu.convert_pd_to_opencv(ids,
object_points,
image_points)
self.video_data.push(image,
ids,
object_points,
image_points)
self.tracking_data.push(device_tracking,
calibration_object_tracking)
number_of_points = image_points.shape[0]
LOGGER.info("Grabbed: Returning %s points.", str(number_of_points))
return number_of_points
[docs] def calibrate(self, flags=0):
"""
Do the video calibration, returning RMS re-projection error.
:param flags: OpenCV calibration flags, eg. cv2.CALIB_FIX_ASPECT_RATIO
:return: RMS projection
"""
rms_proj_err, camera_matrix, dist_coeffs, rvecs, tvecs = \
vc.mono_video_calibration(
self.video_data.object_points_arrays,
self.video_data.image_points_arrays,
(self.video_data.images_array[0].shape[1],
self.video_data.images_array[0].shape[0]),
flags
)
self.calibration_params.set_data(camera_matrix,
dist_coeffs,
rvecs,
tvecs)
LOGGER.info("Mono calibration: rms_proj_err=%s.", str(rms_proj_err))
return rms_proj_err, copy.deepcopy(self.calibration_params)
[docs] def iterative_calibration(self,
number_of_iterations: int,
reference_ids,
reference_image_points,
reference_image_size,
flags: int = 0):
"""
Does iterative calibration, like Datta 2009,
returning RMS re-projection error.
:return: RMS projection
"""
rms_proj_err, param_copy = self.calibrate(flags=flags)
cached_images = copy.deepcopy(self.video_data.images_array)
for i in range(0, number_of_iterations):
images = copy.deepcopy(cached_images)
cu.detect_points_in_canonical_space(
self.point_detector,
self.minimum_points_per_frame,
self.video_data,
images,
self.calibration_params.camera_matrix,
self.calibration_params.dist_coeffs,
reference_ids,
reference_image_points,
reference_image_size)
rms_proj_err, param_copy = self.calibrate(flags=flags)
self.point_detector.set_camera_parameters(
self.calibration_params.camera_matrix,
self.calibration_params.dist_coeffs)
LOGGER.info("Iterative calibration: %s: rms_proj_err=%s.",
str(i), str(rms_proj_err))
return rms_proj_err, param_copy
[docs] def handeye_calibration(self,
override_pattern2marker=None,
use_opencv: bool = True,
do_bundle_adjust: bool = False):
"""
Do handeye calibration, returning RMS re-projection error.
Note: This handeye_calibration on this class assumes you are
tracking both the calibration pattern (e.g. chessboard) and the
device (e.g. laparoscope). So, the calibration routines calibrate
for hand2eye and pattern2marker. If you want something more customised,
work with video_calibration_hand_eye.py.
:param override_pattern2marker: If provided a 4x4 pattern2marker
that is taken as constant.
:param use_opencv: If True we use OpenCV based methods, if false,
Guofang Xiao's method.
:param do_bundle_adjust: If True we do an additional bundle adjustment
at the end.
:return: RMS reprojection error
:rtype: float
"""
rms_proj_err, handeye, pattern2marker = \
vc.mono_handeye_calibration(
self.video_data.object_points_arrays,
self.video_data.image_points_arrays,
self.calibration_params.camera_matrix,
self.calibration_params.dist_coeffs,
self.tracking_data.device_tracking_array,
self.tracking_data.calibration_tracking_array,
self.calibration_params.rvecs,
self.calibration_params.tvecs,
override_pattern2marker=override_pattern2marker,
use_opencv=use_opencv,
do_bundle_adjust=do_bundle_adjust
)
self.calibration_params.set_handeye(handeye, pattern2marker)
return rms_proj_err, copy.deepcopy(self.calibration_params)