def __init__(self, defaultQuat, service_name='update_pose_target'):
self.cameraCalib = CameraCalibration()
rospy.wait_for_service(service_name, timeout=None)
self._update_target = rospy.ServiceProxy(service_name, TrackPose)
self.pose_target_listener = None
self.mouth_target_listener = None
self.defaultQuat = defaultQuat
def test_car_length(self):
cc = CameraCalibration()
X1, Y1 = cc.image2world(661.498, 465.096)
X2, Y2 = cc.image2world(799.19, 617.622)
length = np.sqrt((X2 - X1)**2 + (Y2 - Y1)**2)
real_length = 11
assert np.abs(length - real_length) < 4
def test_car_width(self):
cc = CameraCalibration()
X1, Y1 = cc.image2world(914.768, 603.941)
X2, Y2 = cc.image2world(799.19, 617.622)
width = np.sqrt((X2 - X1)**2 + (Y2 - Y1)**2)
real_width = 2.55
assert np.abs(width - real_width) < 0.2
def calibrate(self):
"""
Compute the camera matrix
"""
self.calib = CameraCalibration(self.nColumns, self.nRows)
self.calib.calibrate(self.srcFolder)
self.params.calib = self.calib
self.nFrames = len(self.calib.srcImgs)
self.stream = ImageListVideoStream(self.calib.srcImgs)
self._setDisplay()
self._setDisplayMax()
self._updateImgProvider()
def __init__(self, filename=join(dirname(dirname(realpath(__file__))), "data/cctv_ftg1_SPEED_000000_carLoc.csv"),
top_obs_rec=906, bottom_obs_rec=439,
window_size=10, start_time_stamp='2018-01-01T08:00'):
self.start_time_stamp = start_time_stamp
self.start_time = []
self.car_type = []
self.track_coord = []
with open(filename, 'r') as f:
for line in f:
dt = line.strip().split("; ")
self.start_time.append(float(dt[0]) * 0.1 / 3)
self.car_type.append(dt[1])
self.track_coord.append([eval(x) for x in dt[2:]])
self.start_time = np.array(self.start_time)
self.car_type = np.array(self.car_type)
self.top_obs_rec = top_obs_rec
self.bottom_obs_rec = bottom_obs_rec
self.window_size = window_size
if pd.to_datetime(self.start_time_stamp) < pd.to_datetime('2018-09-11T16:00'):
self.cc = CameraCalibration(default=False)
self.scale = 1
self.top_obs_rec = 931.761
self.bottom_obs_rec = 439.723
else:
self.cc = CameraCalibration(
p1=(787.919, 640.099), p2=(594.782, 404.630),
p3=(675.065, 399.513), p4=(913.843, 619.450),
u2=703.475, u4=823.241, default=False
)
self.scale = 31.05945702896952 / 51.07394876776243
self.top_obs_rec = 793.968
self.bottom_obs_rec = 337.152
def calibrate(self):
"""
Compute the camera matrix
"""
self.calib = CameraCalibration(self.nColumns, self.nRows)
self.calib.calibrate(self.srcFolder)
self.params.calib = self.calib
self.nFrames = len(self.calib.srcImgs)
self.stream = ImageListVideoStream(self.calib.srcImgs)
self._setDisplay()
self._setDisplayMax()
self._updateImgProvider()
def main():
from glob import glob
from camera_calibration import CameraCalibration
from thresholding import threshold
from transform import PerspectiveTransform
# Fixed image dimensions
height = 720
width = 1280
# Prepare camera calibration
print("Calibrating camera")
calibration = CameraCalibration.default()
# Prepare perspective transform
transform = PerspectiveTransform.default(height, width)
images = glob('test_images/straight*') + glob('test_images/test*')
for fname in images:
print("Processing", fname)
# Run pipeline
img = cv2.imread(fname)
img = calibration.undistort(img)
img, _ = threshold(img)
# img = cv2.morphologyEx(img, cv2.MORPH_CLOSE, (5,5))
img = transform.transform(img)
# Find lines using sliding windows
left_lane, right_lane = find_lines_with_sliding_windows(img)
# Plot sliding windows
plot_sliding_windows(img, left_lane, right_lane)
# combined_binary, color_binary = threshold(img, stack=True)
#
# f, (ax1, ax2, ax3) = plt.subplots(1, 3, figsize=(32, 9))
# ax1.set_title('Stacked thresholds', fontsize=20)
# ax1.imshow(cv2.cvtColor(img, cv2.COLOR_BGR2RGB))
#
# ax2.set_title('Result', fontsize=20)
# ax2.imshow(combined_binary, cmap='gray')
plt.savefig('output_images/sliding_windows_' + fname.split('/')[-1])
def main():
# # Fixed image dimensions
height = 720
width = 1280
transform = PerspectiveTransform.default(height, width)
from camera_calibration import CameraCalibration
calibration = CameraCalibration.default()
images = glob('test_images/straight*') + glob('test_images/test*')
for fname in images:
print("Processing", fname)
image = cv2.cvtColor(cv2.imread(fname), cv2.COLOR_BGR2RGB)
image = calibration.undistort(image)
polygonned = cv2.polylines(np.copy(image), [np.int32(transform.src)],
False,
color=255,
thickness=1)
transformed = transform.transform(np.copy(polygonned))
inversed = transform.invert().transform(np.copy(transformed))
f, (ax1, ax2, ax3, ax4) = plt.subplots(1, 4, figsize=(32, 9))
ax1.set_title('Original', fontsize=20)
ax1.imshow(image)
ax2.set_title('Area of interest', fontsize=20)
ax2.imshow(polygonned)
ax3.set_title('Transformed', fontsize=20)
ax3.imshow(transformed)
ax4.set_title('Transformed inversed', fontsize=20)
ax4.imshow(inversed)
plt.savefig('output_images/transform_' + fname.split('/')[-1])
def main():
# Fixed image dimensions
height = 720
width = 1280
# Prepare camera calibration
print("Calibrating camera")
calibration = CameraCalibration.default()
transform = PerspectiveTransform.default(height, width)
images = glob('test_images/straight*') + glob('test_images/test*')
for fname in images:
print("Processing", fname)
org_image = cv2.imread(fname)
if org_image.shape != (height, width, 3):
print("skipping image", fname, "invalid dimensions", org_image.shape)
continue
# Run the pipeline on a copy of the image
undistorted = calibration.undistort(np.copy(org_image))
transformed = transform.transform(undistorted)
warped_binary, _ = threshold(transformed, stack=False)
# combined, _ = threshold(undistorted, stack=False)
# warped_binary = transform.transform(combined)
lane = Lane()
lane.update(warped_binary)
final = lane.overlay(org_image, draw_lines=False, fill_lane=True, transform=transform.invert())
final = lane.overlay_text(final)
final = cv2.polylines(final, [np.int32(transform.src)], color=[255, 0, 0], isClosed=False)
cv2.imwrite('output_images/lane_{}'.format(fname.split('/')[-1]), final)
final = lane.overlay(np.dstack((warped_binary, warped_binary, warped_binary)) * 255, draw_lines=True,
fill_lane=False)
cv2.imwrite('output_images/lane_warped_{}'.format(fname.split('/')[-1]), final)
class CalibrationIface(PlayerInterface):
"""
Implements the PlayerInterface class with an ImageListVideoStream
It uses the CameraCalibration class to compute the camera matrix from a set of images containing a
chessboard pattern.
"""
def __init__(self, app, context, parent, params, displayName, providerName, timerSpeed=200):
PlayerInterface.__init__(self, app, context, parent, params, displayName, providerName, timerSpeed)
self.nColumns = 9
self.nRows = 6
self.matrixType = 'normal'
self._setDisplay()
self._setDisplay()
@pyqtSlot()
def calibrate(self):
"""
Compute the camera matrix
"""
self.calib = CameraCalibration(self.nColumns, self.nRows)
self.calib.calibrate(self.srcFolder)
self.params.calib = self.calib
self.nFrames = len(self.calib.srcImgs)
self.stream = ImageListVideoStream(self.calib.srcImgs)
self._setDisplay()
self._setDisplayMax()
self._updateImgProvider()
@pyqtSlot(result=QVariant)
def getNRows(self):
"""
Get the number of inner rows in the chessboard pattern
"""
return self.nRows
@pyqtSlot(QVariant)
def setNRows(self, nRows):
"""
Set the number of inner rows in the chessboard pattern
"""
self.nRows = int(nRows)
@pyqtSlot(result=QVariant)
def getNColumns(self):
"""
Get the number of inner columns in the chessboard pattern
"""
return self.nColumns
@pyqtSlot(QVariant)
def setNColumns(self, nColumns):
"""
Set the number of inner rows in the chessboard pattern
"""
self.nColumns = int(nColumns)
@pyqtSlot(result=QVariant)
def getFolderPath(self):
"""
Get the path to the folder where the images with the pattern are stored
"""
diag = QFileDialog()
srcFolder = diag.getExistingDirectory(parent=diag, caption="Chose directory",
directory=os.getenv('HOME'))
self.srcFolder = srcFolder
return srcFolder
@pyqtSlot(QVariant)
def setMatrixType(self, matrixType):
"""
Set the matrix type to be saved. Resolution independant (normal) or dependant (optimized)
:param string matrixType: The type of matrix to be saved. One of ['normal', 'optimized']
"""
matrixType = matrixType.lower()
if matrixType not in ['normal', 'optimized']:
raise KeyError("Expected one of ['normal', 'optimized'], got {}".format(matrixType))
else:
self.matrixType = matrixType
@pyqtSlot()
def saveCameraMatrix(self):
"""
Save the camera matrix selected as self.matrixType
"""
diag = QFileDialog()
destPath = diag.getSaveFileName(parent=diag,
caption='Save matrix',
directory=os.getenv('HOME'),
filter='Numpy (.npy)')
destPath = destPath[0]
if destPath:
if self.matrixType == 'normal':
np.save(destPath, self.cameraMatrix)
elif self.matrixType == 'optimized':
np.save(destPath, self.optimalCameraMatrix)
@pyqtSlot(QVariant)
def setFrameType(self, frameType):
"""
Selects the type of frame to be displayed. (Before, during or after distortion correction)
:param string frameType: The selected frame type. One of ['source', 'detected', 'corrected']
"""
frameType = frameType.lower()
currentIndex = self.stream.currentFrameIdx
if frameType == "source":
imgs = self.calib.srcImgs
elif frameType == "detected":
imgs = self.calib.detectedImgs
elif frameType == "corrected":
imgs = self.calib.correctedImgs
else:
raise KeyError("Expected one of ['source', 'detected', 'corrected'], got {}".format(frameType))
self.stream = ImageListVideoStream(imgs)
self._updateImgProvider()
self.stream.currentFrameIdx = self._validateFrameIdx(currentIndex -1) # reset to previous position
self.getImg()
import cv2
from camera_calibration import CameraCalibration
import image_processing as imgproc
import numpy as np
from line import Line
scaled_size = 1
image_size = (int(1280 * scaled_size), int(720 * scaled_size))
offset = image_size[1] * 0.3
file_name = './output_images/undist_straight_lines1.jpg'
calibration = CameraCalibration('camera_cal/')
calibration.calibrate()
intrinsic_mat = calibration.get_intrinsic()
dist_paras = calibration.get_distortion_paras()
perspective_src_points = scaled_size * np.float32(
[[233, 694], [595, 450], [686, 450], [1073, 694]]) # These points are manually selected
perspective_dst_points = np.float32([[offset, image_size[1]], [offset, 0],
[image_size[0] - offset, 0], [image_size[0] - offset, image_size[1]]])
#######################
img = cv2.imread(file_name)
undist = calibration.distort_correction(img)
undist = cv2.resize(undist, (0, 0), fx=scaled_size, fy=scaled_size)
hls = cv2.cvtColor(undist, cv2.COLOR_BGR2HSV)
h = hls[:, :, 0]
s = hls[:, :, 1]
v = hls[:, :, 2]
def create_camera_calibration():
return CameraCalibration.default()
class CalculateSpeed(object):
def __init__(self, filename=join(dirname(dirname(realpath(__file__))), "data/cctv_ftg1_SPEED_000000_carLoc.csv"),
top_obs_rec=906, bottom_obs_rec=439,
window_size=10, start_time_stamp='2018-01-01T08:00'):
self.start_time_stamp = start_time_stamp
self.start_time = []
self.car_type = []
self.track_coord = []
with open(filename, 'r') as f:
for line in f:
dt = line.strip().split("; ")
self.start_time.append(float(dt[0]) * 0.1 / 3)
self.car_type.append(dt[1])
self.track_coord.append([eval(x) for x in dt[2:]])
self.start_time = np.array(self.start_time)
self.car_type = np.array(self.car_type)
self.top_obs_rec = top_obs_rec
self.bottom_obs_rec = bottom_obs_rec
self.window_size = window_size
if pd.to_datetime(self.start_time_stamp) < pd.to_datetime('2018-09-11T16:00'):
self.cc = CameraCalibration(default=False)
self.scale = 1
self.top_obs_rec = 931.761
self.bottom_obs_rec = 439.723
else:
self.cc = CameraCalibration(
p1=(787.919, 640.099), p2=(594.782, 404.630),
p3=(675.065, 399.513), p4=(913.843, 619.450),
u2=703.475, u4=823.241, default=False
)
self.scale = 31.05945702896952 / 51.07394876776243
self.top_obs_rec = 793.968
self.bottom_obs_rec = 337.152
def coord_image2word(self):
self.Us = []
self.Vs = []
self.times = []
for i, track in enumerate(self.track_coord):
dt = np.dtype('float,float')
track = np.array(track, dtype=dt)
U = track['f0']
V = track['f1']
time = self.start_time[i] + np.arange(U.shape[0]) * 0.1
idx = ((V >= self.bottom_obs_rec) & (V <= self.top_obs_rec))
U = U[idx]
V = V[idx]
time = time[idx]
self.Us.append(U)
self.Vs.append(V)
self.times.append(time)
self.Xs = []
self.Ys = []
for U, V in zip(self.Us, self.Vs):
X, Y = self.cc.image2world(U, V)
self.Xs.append(X)
self.Ys.append(Y)
return self.Xs, self.Ys, self.times
def get_speed(self):
Xs, Ys, times = self.coord_image2word()
self.speeds = []
for X, Y, time in zip(Xs, Ys, times):
# self.speeds.append(
# ((np.sqrt(
# (X[self.window_size:] - X[:-self.window_size])**2 +
# (Y[self.window_size:] - Y[:-self.window_size])**2
# ) / (time[self.window_size:] - time[:-self.window_size]) * 3.6)).mean()
# )
if len(time) > 1:
self.speeds.append(np.sqrt((X[-1] - X[0])**2 + (Y[-1] - Y[0])**2) / (time[-1] - time[0]) * 3.6 * self.scale)
# m/s -> km/h
return self.speeds
def get_df(self, save=False):
df = pd.DataFrame(columns=['time', 'speed', 'car_type'])
# all_speeds = np.hstack(self.speeds)
all_speeds = self.speeds
# all_times = [(tm[self.window_size:] + tm[:-self.window_size]) / 2 for tm in self.times]
all_times = [(tm[-1] + tm[0]) / 2 for tm in self.times if len(tm) > 1]
all_times = np.hstack(all_times).astype('timedelta64[s]') + np.datetime64(self.start_time_stamp)
df.speed = all_speeds
df.time = all_times
i = 0
for j in range(len(self.times)):
# df.loc[i: i+self.speeds[j].shape[0], 'car_type'] = self.car_type[j]
# i = i + self.speeds[j].shape[0]
if len(self.times[j]) > 1:
df.loc[i, 'car_type'] = self.car_type[j]
i += 1
if save:
df.to_csv(join(dirname(dirname(realpath(__file__))), f'result/speeds_{self.start_time_stamp}.csv'), index=False)
return df
def draw_result(self):
for i, (car_type, speed, time) in enumerate(zip(self.car_type, self.speeds, self.times)):
plt.close()
plt.plot((time[self.window_size:] + time[:-self.window_size]) / 2, speed, 'o-')
plt.title(f"{car_type} speed")
plt.xlabel('time (s)')
plt.ylabel('speed (km/h)')
plt.savefig(join(dirname(dirname(realpath(__file__))), f'result/speed/{i}.pdf'))
plt.close()
for i, (car_type, X, Y) in enumerate(zip(self.car_type, self.Xs, self.Ys)):
plt.close()
plt.plot(X, Y)
plt.title(f"{car_type} track")
plt.xlabel('X (m)')
plt.ylabel('Y (m)')
plt.savefig(join(dirname(dirname(realpath(__file__))), f'result/track/{i}.pdf'))
plt.close()
for i, (car_type, u, v) in enumerate(zip(self.car_type, self.Us, self.Vs)):
plt.close()
plt.plot(u, v)
plt.title(f"{car_type} track")
plt.xlabel('U (pixel)')
plt.ylabel('V (pixel)')
plt.savefig(join(dirname(dirname(realpath(__file__))),f'result/track_image/{i}.pdf'))
plt.close()
return 0
left_fit, right_fit = polyfit.fit_polynomial(left_fit_prev,
right_fit_prev)
# Warp image with lanes back
out_image = transform.warp_back(undist, polyfit)
# Curvature and offset
curvature = polyfit.curvature()
offset = polyfit.offset()
return out_image, left_fit, right_fit, curvature, offset
if __name__ == "__main__":
# Undistortion parameters
undis_param = CameraCalibration.load_pickle_with_undist_params(
"camera_cal/wide_dist_pickle.p")
# Threshold parameters
ksize = 3 # Sobel kernel size
s_thresh_img = (170, 255)
s_thresh_vid = (80, 200)
sx_thresh = (20, 200)
sy_thresh = (20, 200)
m_thresh = (30, 120)
d_thresh = (0.7, 1.3)
# Input/output folders of test images
in_folder = os.getcwd() + "/test_images/"
out_folder = os.getcwd() + "/output_images/"
# Detect lanes on test images and write to out folder
def scaled_sobel(abs_sobel):
return np.uint8(255 * abs_sobel / np.max(abs_sobel))
@staticmethod
def create_binary(image):
return np.zeros_like(image)
@staticmethod
def region_of_interest(img, vertices):
mask = np.zeros_like(img)
if len(img.shape) > 2:
channel_count = img.shape[2]
ignore_mask_color = (255, ) * channel_count
else:
ignore_mask_color = 255
cv2.fillPoly(mask, vertices, ignore_mask_color)
masked_image = cv2.bitwise_and(img, mask)
return masked_image
run = False
if run:
camcal = CameraCalibration()
camcal.calibrate_camera()
det = LaneLineDetector(camcal)
output = 'harder_challenge_video_submit.mp4'
clip1 = VideoFileClip('harder_challenge_video.mp4')
clip = clip1.fl_image(det.pipeline)
clip.write_videofile(output, audio=False)
class CalibrationIface(PlayerInterface):
"""
Implements the PlayerInterface class with an ImageListVideoStream
It uses the CameraCalibration class to compute the camera matrix from a set of images containing a
chessboard pattern.
"""
def __init__(self,
app,
context,
parent,
params,
displayName,
providerName,
timerSpeed=200):
PlayerInterface.__init__(self, app, context, parent, params,
displayName, providerName, timerSpeed)
self.nColumns = 9
self.nRows = 6
self.matrixType = 'normal'
self._setDisplay()
self._setDisplay()
@pyqtSlot()
def calibrate(self):
"""
Compute the camera matrix
"""
self.calib = CameraCalibration(self.nColumns, self.nRows)
self.calib.calibrate(self.srcFolder)
self.params.calib = self.calib
self.nFrames = len(self.calib.srcImgs)
self.stream = ImageListVideoStream(self.calib.srcImgs)
self._setDisplay()
self._setDisplayMax()
self._updateImgProvider()
@pyqtSlot(result=QVariant)
def getNRows(self):
"""
Get the number of inner rows in the chessboard pattern
"""
return self.nRows
@pyqtSlot(QVariant)
def setNRows(self, nRows):
"""
Set the number of inner rows in the chessboard pattern
"""
self.nRows = int(nRows)
@pyqtSlot(result=QVariant)
def getNColumns(self):
"""
Get the number of inner columns in the chessboard pattern
"""
return self.nColumns
@pyqtSlot(QVariant)
def setNColumns(self, nColumns):
"""
Set the number of inner rows in the chessboard pattern
"""
self.nColumns = int(nColumns)
@pyqtSlot(result=QVariant)
def getFolderPath(self):
"""
Get the path to the folder where the images with the pattern are stored
"""
diag = QFileDialog()
srcFolder = diag.getExistingDirectory(parent=diag,
caption="Chose directory",
directory=os.getenv('HOME'))
self.srcFolder = srcFolder
return srcFolder
@pyqtSlot(QVariant)
def setMatrixType(self, matrixType):
"""
Set the matrix type to be saved. Resolution independant (normal) or dependant (optimized)
:param string matrixType: The type of matrix to be saved. One of ['normal', 'optimized']
"""
matrixType = matrixType.lower()
if matrixType not in ['normal', 'optimized']:
raise KeyError(
"Expected one of ['normal', 'optimized'], got {}".format(
matrixType))
else:
self.matrixType = matrixType
@pyqtSlot()
def saveCameraMatrix(self):
"""
Save the camera matrix selected as self.matrixType
"""
diag = QFileDialog()
destPath = diag.getSaveFileName(parent=diag,
caption='Save matrix',
directory=os.getenv('HOME'),
filter='Numpy (.npy)')
destPath = destPath[0]
if destPath:
if self.matrixType == 'normal':
np.save(destPath, self.cameraMatrix)
elif self.matrixType == 'optimized':
np.save(destPath, self.optimalCameraMatrix)
@pyqtSlot(QVariant)
def setFrameType(self, frameType):
"""
Selects the type of frame to be displayed. (Before, during or after distortion correction)
:param string frameType: The selected frame type. One of ['source', 'detected', 'corrected']
"""
frameType = frameType.lower()
currentIndex = self.stream.currentFrameIdx
if frameType == "source":
imgs = self.calib.srcImgs
elif frameType == "detected":
imgs = self.calib.detectedImgs
elif frameType == "corrected":
imgs = self.calib.correctedImgs
else:
raise KeyError(
"Expected one of ['source', 'detected', 'corrected'], got {}".
format(frameType))
self.stream = ImageListVideoStream(imgs)
self._updateImgProvider()
self.stream.currentFrameIdx = self._validateFrameIdx(
currentIndex - 1) # reset to previous position
self.getImg()
class TrackerInterface:
def __init__(self, defaultQuat, service_name='update_pose_target'):
self.cameraCalib = CameraCalibration()
rospy.wait_for_service(service_name, timeout=None)
self._update_target = rospy.ServiceProxy(service_name, TrackPose)
self.pose_target_listener = None
self.mouth_target_listener = None
self.defaultQuat = defaultQuat
#### PUBLIC METHODS
# we try to make it so that all of these methods are idempotent
# and can be called from any state
def start_updating_target_to_pose(self,
target_pose_topic,
robot_coord_offset=[0, 0, 0]):
self.stop_moving()
self.pose_target_listener = rospy.Subscriber(
target_pose_topic, Pose, self._update_target_pose_robot_frame,
(robot_coord_offset))
def start_updating_target_to_point(self,
mouth_point_topic,
robot_coord_offset=[0, 0, 0]):
self.stop_moving()
self.mouth_target_listener = rospy.Subscriber(
mouth_point_topic, PointStamped, self._update_target_camera_frame,
(robot_coord_offset))
def start_tracking_fixed_pose(self, robot_coord_point, robot_coord_quat):
self.stop_moving()
# you just send the target point, you don't need to continually update it
self._update_target(target=Pose(
Point(robot_coord_point[0], robot_coord_point[1],
robot_coord_point[2]), robot_coord_quat))
def start_tracking_fixed_target(self, robot_coord_point):
self.start_tracking_fixed_pose(robot_coord_point, self.defaultQuat)
def stop_moving(self):
self._stop_updating_target()
self._update_target(stopMotion=True)
#### PRIVATE METHODS #####
def _stop_updating_target(self):
if (self.mouth_target_listener is not None):
self.mouth_target_listener.unregister()
if (self.pose_target_listener is not None):
self.pose_target_listener.unregister()
# return an np.array of the [x,y,z] target mouth location
# in the coordinate frame of the robot base
def _convert_camera_to_robot_frame(self, mouth_pos):
t = mouth_pos.point
point_in_camera_frame = np.array([t.x, t.y, t.z, 1])
point_in_robot_frame = self.cameraCalib.convert_to_robot_frame(
point_in_camera_frame)
return point_in_robot_frame[0:3]
# compute and move toward the target mouth location
# only move for at most timeoutSecs,
def _update_target_camera_frame(self,
mouth_pos,
robot_coord_offset=[0, 0, 0]):
endLoc = self._convert_camera_to_robot_frame(mouth_pos) + np.array(
robot_coord_offset)
self._update_target(target=Pose(Point(endLoc[0], endLoc[1], endLoc[2]),
self.defaultQuat))
# move toward the target spoon pose
def _update_target_pose_robot_frame(self,
target_pose,
robot_coord_offset=[0, 0, 0]):
newPosition = Point(target_pose.position.x + robot_coord_offset[0],
target_pose.position.y + robot_coord_offset[1],
target_pose.position.z + robot_coord_offset[2])
self._update_target(target=Pose(newPosition, target_pose.orientation))
grad_color_stack, img = color_gradient_threshold_pipeline(image, img_processor)
birds_eye, M, Minv = calibration.perspective_transform(img, src, dst)
undist = calibration.undistort(image)
search, left_fitx, right_fitx, ploty, image_final = find_lane_lines.find(birds_eye, undist, Minv, left_line, right_line)
return image_final
#__main__
# Camera Calibration
img = mpimg.imread('camera_cal/calibration1.jpg')
img_size = (img.shape[1], img.shape[0])
calibration = CameraCalibration(chessboard_images_dir='camera_cal', pattern_size=(9,6))
calibration.calibrate_camera(img_size)
# Image Processor
img_processor = ImageProcessor()
# src and dst for perspective transform
test_img = mpimg.imread('test_images/straight_lines2.jpg')
test_img_size = (test_img.shape[1], test_img.shape[0])
src = np.float32([[580,460], [710,460], [1150,720], [150,720]])
offset = 200
dst = np.float32([ [offset, 0],
[test_img_size[0]-offset, 0],
[test_img_size[0]-offset, test_img_size[1]-0],
[offset, test_img_size[1]-0]
continue
gtDistances += cv_dists
tofDistances += tof_dists
stereoDistances += stereo_dists
# print("-----image ", fname, "-----")
# for i in range(len(cv_dists)):
# print('corner {:02d} has gt = {:.1f}, tof = {:.1f}, stereo = {:.1f}'.format(
# i, cv_dists[i], tof_dists[i], stereo_dists[i]))
return [gtDistances, tofDistances, stereoDistances]
if __name__ == '__main__':
square_size = 26.3571
width = 6
height = 5
cal_obj = CameraCalibration(square_size, width, height)
rgbModel = 'rgbCamera.yml'
tofModel = 'irCamera.yml'
stereoModel = 'stereo.yml'
cal_obj.load_cameraModel(rgbModel, tofModel, stereoModel)
dirpath = '../data/bias_calibration/corner*'
gtDistances, tofs, stereo_distances = calibrateDepth(
cal_obj,
dirpath,
irPrefix='Gray_',
depthPrefix='DepthImage_',
rgbPrefix='RBG_',
ifVisualization=False,
ifDivide16=True,
ifUseRGB=True)
