def test_on_video_with_calib():
"""
connect track and close with single tag,
and calibrated camera
reqs: 03, 04 ,05
"""
config1 = {
'video source': 'data/output.avi',
'calibration': 'data/calibration.txt',
'marker size': 50,
}
tracker = ArUcoTracker(config1)
tracker.start_tracking()
for frame in range(10):
(port_handles, _timestamps, _framenumbers, tracking,
_quality) = tracker.get_frame()
if frame == 1:
regression_array = np.array(
[[0.99928828, -0.02261964, 0.03018762, 1.7609988e+01],
[-0.02280416, -0.99972323, 0.00578211, 2.5085676e+01],
[0.03004847, -0.00646639, -0.99952753, 2.1036779e+02],
[0., 0., 0., 1.]])
assert np.allclose(tracking[port_handles.index('DICT_4X4_50:0')],
regression_array)
tracker.stop_tracking()
tracker.close()
def test_on_static_muti_tag():
"""
connect track and close with multi tag,
reqs: 03, 04 ,05, 07
"""
config = {
'video source': 'data/12markers.avi',
'aruco dictionary': 'DICT_6X6_250'
}
tracker = ArUcoTracker(config)
tracker.start_tracking()
(port_handles, timestamps, framenumbers, tracking,
quality) = tracker.get_frame()
assert len(port_handles) == len(timestamps)
assert len(port_handles) == len(framenumbers)
assert len(port_handles) == len(tracking)
assert len(port_handles) == len(quality)
assert len(port_handles) == 12
for tagid in range(1, 13):
tag_name = str('DICT_6X6_250:' + str(tagid))
assert tag_name in port_handles
regression_array6 = np.array([[1., 0., 0., 262.5], [0., 1., 0., 241.5],
[0., 0., 1., -151.32085], [0., 0., 0., 1.]])
assert np.allclose(tracking[port_handles.index('DICT_6X6_250:6')],
regression_array6)
tracker.stop_tracking()
tracker.close()
def test_on_video_with_debug():
"""
connect track and close with single tag, with debug flag on
debug should open a separate window showing the image capture
reqs: 03, 04 ,05
"""
config = {'video source': 'data/output.avi', 'debug': True}
tracker = ArUcoTracker(config)
tracker.start_tracking()
for frame in range(2):
(port_handles, timestamps, framenumbers, tracking,
quality) = tracker.get_frame()
assert len(port_handles) == len(timestamps)
assert len(port_handles) == len(framenumbers)
assert len(port_handles) == len(tracking)
assert len(port_handles) == len(quality)
assert len(port_handles) == 1
assert port_handles[0] == 'DICT_4X4_50:0'
assert framenumbers[0] == frame
assert quality[0] == 1.0
tracker.stop_tracking()
tracker.close()
def test_no_video_single_tag():
"""
raises a value error when no video and no image passed.
works on static images
reqs: 03, 04 ,05
"""
config = {'video source': 'none'}
tracker = ArUcoTracker(config)
assert not tracker.has_capture()
tracker.start_tracking()
with pytest.raises(ValueError):
tracker.get_frame()
capture = VideoCapture('data/output.avi')
for frame in range(2):
_, image = capture.read()
(port_handles, timestamps, framenumbers, tracking,
quality) = tracker.get_frame(image)
assert len(port_handles) == len(timestamps)
assert len(port_handles) == len(framenumbers)
assert len(port_handles) == len(tracking)
assert len(port_handles) == len(quality)
assert len(port_handles) == 1
assert port_handles[0] == 'DICT_4X4_50:0'
assert framenumbers[0] == frame
assert quality[0] == 1.0
tracker.stop_tracking()
tracker.close()
def test_with_tool_desc_and_calib():
"""
connect track and close with multitags, defined
rigid bodies, and camera calibration
reqs: 03, 04 ,05, 07
"""
config = {
'video source':
'data/multipattern.avi',
'calibration':
'data/calibration.txt',
'rigid bodies': [{
'name': 'reference',
'filename': 'data/reference.txt',
'aruco dictionary': 'DICT_ARUCO_ORIGINAL',
'tag width': 49.50
}, {
'name': 'pointer',
'filename': 'data/pointer.txt',
'aruco dictionary': 'DICT_ARUCO_ORIGINAL'
}]
}
tracker = ArUcoTracker(config)
tracker.start_tracking()
assert tracker.has_capture()
(port_handles, timestamps, framenumbers, tracking,
quality) = tracker.get_frame()
assert len(port_handles) == len(timestamps)
assert len(port_handles) == len(framenumbers)
assert len(port_handles) == len(tracking)
assert len(port_handles) == len(quality)
assert len(port_handles) == 4 #there is an extraneous marker (1000)
assert 'reference' in port_handles
assert 'pointer' in port_handles
assert 'DICT_4X4_50:0' in port_handles
reference_index = port_handles.index('reference')
pointer_index = port_handles.index('pointer')
assert np.isclose(quality[reference_index], 0.91666666)
assert np.isclose(quality[pointer_index], 0.83333333)
ref_regression = np.array(
[[-0.84701057, 0.51884094, -0.11565978, -8.22903442e+01],
[-0.48733129, -0.67100208, 0.55880625, 4.85032501e+01],
[0.21232361, 0.52967943, 0.82119327, 2.43992401e+02],
[0.00000000e+00, 0.00000000e+00, 0.00000000e+00, 1.00000000e+00]],
dtype=np.float32)
assert np.allclose(tracking[reference_index], ref_regression)
tracker.stop_tracking()
tracker.close()
def test_iteration_over_empty_dicts():
"""
Tests that when no markers are detected in an early
dictionary, we continue to iterate over the remainder
issue #29
"""
config = {
'video source':
'data/multipattern.avi',
'rigid bodies': [
{
'name': 'bad_pointer',
'filename': 'data/pointer.txt',
'aruco dictionary': 'DICT_7X7_250'
},
{
'name': 'reference',
'filename': 'data/reference.txt',
'aruco dictionary': 'DICT_ARUCO_ORIGINAL'
},
{
'name': 'good_pointer',
'filename': 'data/pointer.txt',
'aruco dictionary': 'DICT_ARUCO_ORIGINAL'
},
]
}
tracker = ArUcoTracker(config)
tracker.start_tracking()
(port_handles, timestamps, framenumbers, tracking,
quality) = tracker.get_frame()
print(port_handles)
assert len(port_handles) == len(timestamps)
assert len(port_handles) == len(framenumbers)
assert len(port_handles) == len(tracking)
assert len(port_handles) == len(quality)
assert len(port_handles) == 5
assert 'reference' in port_handles
assert 'bad_pointer' in port_handles
assert 'good_pointer' in port_handles
assert 'DICT_4X4_50:0' in port_handles
ref_regression = np.array([[1., 0., 0., 135.38637], [0., 1., 0., 272.5],
[0., 0., 1., -57.1915], [0., 0., 0., 1.]],
dtype=np.float32)
pointer_index = port_handles.index('bad_pointer')
assert np.all(np.isnan(tracking[pointer_index][0:3, 0:4]))
reference_index = port_handles.index('reference')
assert np.allclose(tracking[reference_index], ref_regression)
def test_start_tracking_throws():
"""
Tests that value error is thrown when start tracking called when not ready.
reqs:
"""
config = {'video source': 'data/12markers.avi'}
tracker = ArUcoTracker(config)
tracker.start_tracking()
with pytest.raises(ValueError):
tracker.start_tracking()
tracker.close()
def test_with_no_tags_and_calib():
"""
connect track and close with multitags, defined
rigid bodies, but no tags
visible, should return an NaN matrix for each
undetected body
reqs: 03, 04 ,05, 07
"""
config = {
'calibration':
'data/calibration.txt',
'video source':
'none',
'rigid bodies': [{
'name': 'reference',
'filename': 'data/reference.txt',
'aruco dictionary': 'DICT_4X4_50'
}, {
'name': 'pointer',
'filename': 'data/pointer.txt',
'aruco dictionary': 'DICT_4X4_50'
}]
}
tracker = ArUcoTracker(config)
tracker.start_tracking()
image = np.zeros((640, 480, 3), dtype=np.uint8)
(port_handles, timestamps, framenumbers, tracking,
quality) = tracker.get_frame(image)
assert len(port_handles) == len(timestamps)
assert len(port_handles) == len(framenumbers)
assert len(port_handles) == len(tracking)
assert len(port_handles) == len(quality)
assert len(port_handles) == 2
assert 'reference' in port_handles
assert 'pointer' in port_handles
reference_index = port_handles.index('reference')
pointer_index = port_handles.index('pointer')
assert quality[reference_index] == 0
assert quality[pointer_index] == 0
assert np.all(np.isnan(tracking[reference_index][0:3, 0:4]))
assert np.all(np.isnan(tracking[pointer_index][0:3, 0:4]))
tracker.stop_tracking()
tracker.close()
def run_demo(tracked, dicom_dir):
""" Demo """
app = QtWidgets.QApplication([])
if tracked:
tracker = ArUcoTracker({})
tracker.start_tracking()
slice_viewer = TrackedSliceViewer(dicom_dir, tracker)
else:
slice_viewer = MouseWheelSliceViewer(dicom_dir)
slice_viewer.start()
app.exec_()
def test_no_video_no_tag():
"""
no errors when no tags detected
works on static images
reqs: 03, 04 ,05
"""
config = {'video source': 'none'}
tracker = ArUcoTracker(config)
tracker.start_tracking()
image = np.zeros((640, 480, 3), dtype=np.uint8)
(port_handles, timestamps, framenumbers, tracking,
quality) = tracker.get_frame(image)
assert len(port_handles) == len(timestamps)
assert len(port_handles) == len(framenumbers)
assert len(port_handles) == len(tracking)
assert len(port_handles) == len(quality)
assert len(port_handles) == 0
tracker.stop_tracking()
tracker.close()
def test_setting_capture_properties():
"""
connect track and close with multi tag,
reqs: 03, 04 ,05, 07
"""
config = {
'video source': 'data/12markers.avi',
'aruco dictionary': 'DICT_6X6_250',
"capture properties": {
"CAP_PROP_FRAME_WIDTH": 1280,
"CAP_PROP_FRAME_HEIGHT": 1024
}
}
tracker = ArUcoTracker(config)
tracker.start_tracking()
(_port_handles, _timestamps, _framenumbers, _tracking,
_quality) = tracker.get_frame()
tracker.stop_tracking()
tracker.close()
def test_on_video_with_single_tag():
"""
connect track and close with single tag,
reqs: 03, 04 ,05
"""
config = {'video source': 'data/output.avi'}
tracker = ArUcoTracker(config)
tracker.start_tracking()
for frame in range(10):
(port_handles, timestamps, framenumbers, tracking,
quality) = tracker.get_frame()
assert len(port_handles) == len(timestamps)
assert len(port_handles) == len(framenumbers)
assert len(port_handles) == len(tracking)
assert len(port_handles) == len(quality)
assert len(port_handles) == 1
assert port_handles[0] == 'DICT_4X4_50:0'
assert framenumbers[0] == frame
assert quality[0] == 1.0
tracker.stop_tracking()
tracker.close()
def test_vs_solve_pnp_singletag():
"""
This checks whether the tracking result using skarucotracker
matches that when using the solvePnP approach as originally
implemented by scikit-surgeryBARAD, with the addition of
tracking smoothing
"""
calib_mtx = np.array([[608.67179504, 0.00000000, 323.12263928],
[0.00000000, 606.13421375, 231.29247171],
[0.0, 0.0, 1.0]], dtype = np.float64)
distortion = np.array([-0.02191634, -0.14300148, -0.00395124,
-0.00044941, 0.19656551], dtype = np.float64)
videofile = 'data/output.avi'
config = {'video source' : 'none',
'camera projection' : calib_mtx,
'camera distortion' : distortion,
'aruco dictionary' : 'DICT_ARUCO_ORIGINAL',
'smoothing buffer' : 5,
'rigid bodies' : [
{
'name' : 'reference',
'filename' : 'data/tag_0.txt',
'aruco dictionary' : 'DICT_4X4_50'
}
]
}
config2 = {'video source' : 'none',
'camera projection' : calib_mtx,
'camera distortion' : distortion,
'aruco dictionary' : 'DICT_4X4_50',
'marker size' : 33,
'smoothing buffer' : 5,
}
capture = cv.VideoCapture(videofile)
#set up skarucotracker
tracker = ArUcoTracker(config)
tracker.start_tracking()
tracker2 = ArUcoTracker(config2)
tracker2.start_tracking()
#set up solvepnp
three_d_points = np.loadtxt('data/tag_0.txt')
three_d_points = three_d_points.reshape(1, 16)
three_d_points = ccw_to_cw(three_d_points)
reference_register = Registration2D3D(three_d_points,
calib_mtx, distortion,
buffer_size=5)
for _frame in range(10):
_, image = capture.read()
(port_handles, _timestamps, _framenumbers,
tracking, _quality) = tracker.get_frame(image)
assert 'reference' in port_handles
reference_index = port_handles.index('reference')
aruco_reference_tracking = tracking[reference_index]
(port_handles, _timestamps, _framenumbers,
tracking, _quality) = tracker2.get_frame(image)
assert 'DICT_4X4_50:0' in port_handles
reference_index = port_handles.index('DICT_4X4_50:0')
aruco_reference_tracking2= tracking[reference_index]
#now try again using our own implementation using cv.solvepnp as
#formerly implemented in BARD
marker_corners, ids, _ = aruco.detectMarkers(image,
aruco.getPredefinedDictionary(
aruco.DICT_4X4_50))
success, modelreference2camera = \
reference_register.get_matrix(
ids, marker_corners)
assert success
assert np.allclose(modelreference2camera, aruco_reference_tracking)
assert np.allclose(aruco_reference_tracking2, aruco_reference_tracking)
tracker.stop_tracking()
tracker.close()
capture.release()
def test_arucotracker_vs_solve_pnp():
"""
This checks whether the tracking result using skarucotracker
matches that when using the solvePnP approach as originally
implemented by scikit-surgeryBARD
"""
calib_mtx = np.array([[608.67179504, 0.00000000, 323.12263928],
[0.00000000, 606.13421375, 231.29247171],
[0.0, 0.0, 1.0]], dtype = np.float64)
distortion = np.array([-0.02191634, -0.14300148, -0.00395124,
-0.00044941, 0.19656551], dtype = np.float64)
videofile = 'data/multipattern.avi'
config = {'video source' : 'none',
'camera projection' : calib_mtx,
'camera distortion' : distortion,
'aruco dictionary' : 'DICT_ARUCO_ORIGINAL',
'rigid bodies' : [
{
'name' : 'reference',
'filename' : 'data/reference.txt',
'aruco dictionary' : 'DICT_ARUCO_ORIGINAL'
}
]
}
tracker = ArUcoTracker(config)
tracker.start_tracking()
capture = cv.VideoCapture(videofile)
_, image = capture.read()
(port_handles, _timestamps, _framenumbers,
tracking, quality) = tracker.get_frame(image)
assert 'reference' in port_handles
reference_index = port_handles.index('reference')
assert np.isclose(quality[reference_index], 0.91666666)
ref_regression = np.array([
[-0.85007277, 0.51807849, -0.09471516, -8.06235428e+01],
[-0.46167221, -0.64647658, 0.60739345, 5.07177658e+01],
[ 0.25344635, 0.56005599, 0.78873458, 2.50268387e+02],
[ 0.0000e+00, 0.00000e+00, 0.00000e+00, 1.00000000e+00]],
dtype=np.float32)
aruco_reference_tracking = tracking[reference_index]
assert np.allclose(aruco_reference_tracking, ref_regression)
tracker.stop_tracking()
tracker.close()
#now try again using our own implementation using cv.solvepnp as
#formerly implemented in BARD
three_d_points = np.loadtxt('data/reference.txt')
reference_register = Registration2D3D(three_d_points,
calib_mtx, distortion,
buffer_size=1)
marker_corners, ids, _ = aruco.detectMarkers(image,
aruco.getPredefinedDictionary(
aruco.DICT_ARUCO_ORIGINAL))
success, modelreference2camera = \
reference_register.get_matrix(
ids, marker_corners)
assert success
assert np.allclose(modelreference2camera, ref_regression)
capture.release()
def test_tool_desc_and_float64():
"""
connect track and close with multitags, defined
rigid bodies, and camera calibration, defined with
different data types, issue #31
reqs: 03, 04 ,05, 07
"""
config = {
'video source':
'data/multipattern.avi',
'calibration':
'data/calibration.txt',
'aruco dictionary':
'DICT_ARUCO_ORIGINAL',
'rigid bodies': [
{
'name': 'reference',
'filename': 'data/reference.txt',
'aruco dictionary': 'DICT_ARUCO_ORIGINAL'
},
]
}
tracker = ArUcoTracker(config)
tracker.start_tracking()
(port_handles, _timestamps, _framenumbers, tracking,
quality) = tracker.get_frame()
assert 'reference' in port_handles
reference_index = port_handles.index('reference')
assert np.isclose(quality[reference_index], 0.91666666)
ref_regression = np.array(
[[-0.84701057, 0.51884094, -0.11565978, -8.22903442e+01],
[-0.48733129, -0.67100208, 0.55880625, 4.85032501e+01],
[0.21232361, 0.52967943, 0.82119327, 2.43992401e+02],
[0.00000000e+00, 0.00000000e+00, 0.00000000e+00, 1.00000000e+00]],
dtype=np.float32)
assert np.allclose(tracking[reference_index], ref_regression)
tracker.stop_tracking()
tracker.close()
port_handles = None
tracking = None
quality = None
#again, but this time set calibration and distortion separately
calib_mtx = np.array(
[[560.0, 0.0, 320.0], [0.0, 560.0, 240.0], [0.0, 0.0, 1.0]],
dtype=np.float32)
distortion = np.array([0.1, 0.0, 0.0, 0.0, 0.0], dtype=np.float32)
config = {
'video source':
'data/multipattern.avi',
'camera projection':
calib_mtx,
'camera distortion':
distortion,
'aruco dictionary':
'DICT_ARUCO_ORIGINAL',
'rigid bodies': [
{
'name': 'reference',
'filename': 'data/reference.txt',
'aruco dictionary': 'DICT_ARUCO_ORIGINAL'
},
]
}
tracker = ArUcoTracker(config)
tracker.start_tracking()
(port_handles, _timestamps, _framenumbers, tracking,
quality) = tracker.get_frame()
assert 'reference' in port_handles
reference_index = port_handles.index('reference')
assert np.isclose(quality[reference_index], 0.91666666)
assert np.allclose(tracking[reference_index], ref_regression)
tracker.stop_tracking()
tracker.close()
port_handles = None
tracking = None
quality = None
#again, but this time set calibration and distortion as float64
calib_mtx = np.array(
[[560.0, 0.0, 320.0], [0.0, 560.0, 240.0], [0.0, 0.0, 1.0]],
dtype=np.float64)
distortion = np.array([0.1, 0.0, 0.0, 0.0, 0.0], dtype=np.float64)
config = {
'video source':
'data/multipattern.avi',
'camera projection':
calib_mtx,
'camera distortion':
distortion,
'aruco dictionary':
'DICT_ARUCO_ORIGINAL',
'rigid bodies': [
{
'name': 'reference',
'filename': 'data/reference.txt',
'aruco dictionary': 'DICT_ARUCO_ORIGINAL'
},
]
}
tracker = ArUcoTracker(config)
tracker.start_tracking()
(port_handles, _timestamps, _framenumbers, tracking,
quality) = tracker.get_frame()
assert 'reference' in port_handles
reference_index = port_handles.index('reference')
assert np.isclose(quality[reference_index], 0.91666666)
assert np.allclose(tracking[reference_index], ref_regression)
tracker.stop_tracking()
tracker.close()
port_handles = None
tracking = None
quality = None
#again, but this time set calibration and distortion as float
calib_mtx = np.array(
[[560.0, 0.0, 320.0], [0.0, 560.0, 240.0], [0.0, 0.0, 1.0]],
dtype=float)
distortion = np.array([0.1, 0.0, 0.0, 0.0, 0.0], dtype=float)
config = {
'video source':
'data/multipattern.avi',
'camera projection':
calib_mtx,
'camera distortion':
distortion,
'aruco dictionary':
'DICT_ARUCO_ORIGINAL',
'rigid bodies': [
{
'name': 'reference',
'filename': 'data/reference.txt',
'aruco dictionary': 'DICT_ARUCO_ORIGINAL'
},
]
}
tracker = ArUcoTracker(config)
tracker.start_tracking()
(port_handles, _timestamps, _framenumbers, tracking,
quality) = tracker.get_frame()
assert 'reference' in port_handles
reference_index = port_handles.index('reference')
assert np.isclose(quality[reference_index], 0.91666666)
assert np.allclose(tracking[reference_index], ref_regression)
tracker.stop_tracking()
tracker.close()
def test_with_tool_descriptions():
"""
connect track and close with multitags and defined
rigid bodies,
reqs: 03, 04 ,05, 07
"""
#test first with no tool descriptions
config = {'video source': 'data/multipattern.avi'}
tracker = ArUcoTracker(config)
tracker.start_tracking()
#with nothing set we'll only detect tag ID 0, the others
#are from a different dictionary.
(port_handles, timestamps, framenumbers, tracking,
quality) = tracker.get_frame()
assert len(port_handles) == len(timestamps)
assert len(port_handles) == len(framenumbers)
assert len(port_handles) == len(tracking)
assert len(port_handles) == len(quality)
assert len(port_handles) == 1
assert 'DICT_4X4_50:0' in port_handles
tracker.stop_tracking()
tracker.close()
#try again with the right dictionary for the bard marker tags
config = {
'video source': 'data/multipattern.avi',
'aruco dictionary': 'DICT_ARUCO_ORIGINAL'
}
tracker = ArUcoTracker(config)
tracker.start_tracking()
#with nothing set we'll only detect tag ID 0, the others
#are from a different dictionary.
(port_handles, timestamps, framenumbers, tracking,
quality) = tracker.get_frame()
assert len(port_handles) == len(timestamps)
assert len(port_handles) == len(framenumbers)
assert len(port_handles) == len(tracking)
assert len(port_handles) == len(quality)
assert len(port_handles) == 17
assert 'DICT_ARUCO_ORIGINAL:1' in port_handles
#we should load the tag info and check that all tags are found
tracker.stop_tracking()
tracker.close()
config = {
'video source':
'data/multipattern.avi',
'rigid bodies': [{
'name': 'reference',
'filename': 'data/reference.txt',
'aruco dictionary': 'DICT_ARUCO_ORIGINAL'
}, {
'name': 'pointer',
'filename': 'data/pointer.txt',
'aruco dictionary': 'DICT_ARUCO_ORIGINAL'
}]
}
tracker = ArUcoTracker(config)
tracker.start_tracking()
(port_handles, timestamps, framenumbers, tracking,
quality) = tracker.get_frame()
assert len(port_handles) == len(timestamps)
assert len(port_handles) == len(framenumbers)
assert len(port_handles) == len(tracking)
assert len(port_handles) == len(quality)
assert len(port_handles) == 4 #there is an extraneous marker (1000)
assert 'reference' in port_handles
assert 'pointer' in port_handles
assert 'DICT_4X4_50:0' in port_handles
reference_index = port_handles.index('reference')
ref_regression = np.array([[1., 0., 0., 135.38637], [0., 1., 0., 272.5],
[0., 0., 1., -57.1915], [0., 0., 0., 1.]],
dtype=np.float32)
assert np.allclose(tracking[reference_index], ref_regression)
assert np.isclose(quality[reference_index], 0.91666666)
pointer_index = port_handles.index('pointer')
assert np.isclose(quality[pointer_index], 0.83333333)
tracker.stop_tracking()
tracker.close()
#check that tag 0 is not detected when we use only
#DICT_ARUCO_ORIGINAL
config = {
'video source':
'data/multipattern.avi',
'aruco dictionary':
'DICT_ARUCO_ORIGINAL',
'rigid bodies': [{
'name': 'reference',
'filename': 'data/reference.txt',
'aruco dictionary': 'DICT_ARUCO_ORIGINAL'
}, {
'name': 'pointer',
'filename': 'data/pointer.txt',
'aruco dictionary': 'DICT_ARUCO_ORIGINAL'
}]
}
tracker = ArUcoTracker(config)
tracker.start_tracking()
(port_handles, timestamps, framenumbers, tracking,
quality) = tracker.get_frame()
assert len(port_handles) == len(timestamps)
assert len(port_handles) == len(framenumbers)
assert len(port_handles) == len(tracking)
assert len(port_handles) == len(quality)
assert len(port_handles) == 3 #there is an extraneous marker (1000)
assert 'reference' in port_handles
assert 'pointer' in port_handles
assert 'DICT_4X4_50:0' not in port_handles
class OverlayApp(OverlayBaseApp):
"""Inherits from OverlayBaseApp, and adds methods to
detect aruco tags and move the model to follow."""
def __init__(self, image_source):
"""override the default constructor to set up sksurgeryarucotracker"""
#we'll use SciKit-SurgeryArUcoTracker to estimate the pose of the
#visible ArUco tag relative to the camera. We use a dictionary to
#configure SciKit-SurgeryArUcoTracker
ar_config = {
"tracker type":
"aruco",
#Set to none, to share video source with OverlayBaseApp
"video source":
'none',
"debug":
False,
#the aruco tag dictionary to use. DICT_4X4_50 will work with
#../tags/aruco_4by4_0.pdf
"dictionary":
'DICT_4X4_50',
"marker size":
50, # in mm
#We need a calibrated camera. For now let's just use a
#a hard coded estimate. Maybe you could improve on this.
"camera projection":
numpy.array(
[[560.0, 0.0, 320.0], [0.0, 560.0, 240.0], [0.0, 0.0, 1.0]],
dtype=numpy.float32),
"camera distortion":
numpy.zeros((1, 4), numpy.float32)
}
self.tracker = ArUcoTracker(ar_config)
self.tracker.start_tracking()
#and call the constructor for the base class
if sys.version_info > (3, 0):
super().__init__(image_source)
else:
#super doesn't work the same in py2.7
OverlayBaseApp.__init__(self, image_source)
def update(self):
"""Update the background render with a new frame and
scan for aruco tags"""
_, image = self.video_source.read()
self._aruco_detect_and_follow(image)
#Without the next line the model does not show as the clipping range
#does not change to accommodate model motion. Uncomment it to
#see what happens.
self.vtk_overlay_window.set_camera_state({"ClippingRange": [10, 800]})
self.vtk_overlay_window.set_video_image(image)
self.vtk_overlay_window.Render()
def _aruco_detect_and_follow(self, image):
"""Detect any aruco tags present using sksurgeryarucotracker
"""
#tracker.get_frame(image) returns 5 lists of tracking data.
#we'll only use the tracking matrices (tag2camera)
_port_handles, _timestamps, _frame_numbers, tag2camera, \
_tracking_quality = self.tracker.get_frame(image)
if tag2camera is not None:
#pass the first entry in tag2camera. If you have more than one tag
#visible, you may need to do something cleverer here.
self._move_camera(tag2camera[0])
def _move_camera(self, tag2camera):
"""Internal method to move the rendered models in
some interesting way"""
#SciKit-SurgeryCore has a useful TransformManager that makes
#chaining together and inverting transforms more intuitive.
#We'll just use it to invert a matrix here.
transform_manager = TransformManager()
transform_manager.add("tag2camera", tag2camera)
camera2tag = transform_manager.get("camera2tag")
#Let's move the camera, rather than the model this time.
self.vtk_overlay_window.set_camera_pose(camera2tag)