def calibrate(numcam, num_bots, tcp_ips, tcp_ports, buffer_size, cam_pairs):
CAM_IDS = [x % 10 for x in tcp_ports]
cams_to_be_calib = get_cam_to_be_calibrated(cam_pairs)
print 'cams_to_be_calib', cams_to_be_calib
trackers = {}
for y, x in cams_to_be_calib.items():
print x, y
idx = CAM_IDS.index(y)
z = CamTracker(y, num_bots, tcp_ips[idx], tcp_ports[idx], buffer_size)
trackers[y] = z
for i in trackers.keys():
trackers[i].thread.start()
print cam_pairs
tfs = []
for i, j in cam_pairs:
print "======>", i, j
tf = {}
tf['from_to'] = (i, j)
id_i = int(i)
id_j = int(j)
print id_j, id_i
calib_obj = Calibrate(trackers[id_i], trackers[id_j], 'somefilename')
calib_obj.calibrate()
try:
tf['tf'] = calib_obj.tf
except:
continue
print calib_obj.R
tfs.append(tf)
close_trackers(trackers)
return tfs
def main():
print("Start MatLabEngine:\n ")
eng = matlab.engine.start_matlab()
print("Start Calibrate:\n ")
Calibrate.calibrate(eng)
print("Start e-book reader:\n ")
app = QApplication(sys.argv)
window = MainWindow(eng)
window.show()
qt_return_code = app.exec_()
print('Qt return code:' + str(qt_return_code))
sys.exit(qt_return_code)
class HandyAR:
cap_region_x_begin = 0.5 # start point/total width
cap_region_y_end = 0.8 # start point/total width
threshold = 60 # BINARY threshold
blurValue = 41 # GaussianBlur parameter
bgSubThreshold = 50
def __init__(self):
self.isBgCaptured = 0 # bool, whether the background captured
# Camera
self.camera = cv2.VideoCapture(0)
self.camera.set(10, 200)
self.handDetect = HandDetect()
self.calibrate = Calibrate()
self.demo = None
def run(self):
while self.camera.isOpened():
ret, frame = self.camera.read()
frame = cv2.bilateralFilter(frame, 5, 50, 100) # smoothing filter
frame = cv2.flip(frame, 1) # flip the frame horizontally
cv2.rectangle(
frame, (int(self.cap_region_x_begin * frame.shape[1]), 0),
(frame.shape[1], int(self.cap_region_y_end * frame.shape[0])),
(255, 0, 0), 2)
# half of the original frame for opengl operation
halfframe = frame[0:int(self.cap_region_y_end * frame.shape[0]),
int(self.cap_region_x_begin *
frame.shape[1]):frame.shape[1]]
cv2.imshow('original', frame)
if self.isBgCaptured == 1: # this part wont run until background captured
img = self.handDetect.removeBG(frame)
img = img[0:int(self.cap_region_y_end * frame.shape[0]),
int(self.cap_region_x_begin *
frame.shape[1]):frame.shape[1]] # clip the ROI
#cv2.imshow('mask', img)
# convert image into binary image
gray = cv2.cvtColor(img, cv2.COLOR_BGR2GRAY)
blur = cv2.GaussianBlur(gray, (self.blurValue, self.blurValue),
0)
#cv2.imshow('blur', blur)
ret, thresh = cv2.threshold(blur, self.threshold, 255,
cv2.THRESH_BINARY)
#cv2.imshow('ori', thresh)
# get coutours
thresh1 = copy.deepcopy(thresh)
contours, hierarchy = cv2.findContours(thresh1, cv2.RETR_TREE,
cv2.CHAIN_APPROX_SIMPLE)
length = len(contours)
maxArea = -1
if length > 0:
for i in range(
length
): # find the biggest contour (according to area)
temp = contours[i]
area = cv2.contourArea(temp)
if area > maxArea:
maxArea = area
maxi = i
contour = contours[maxi]
hull = cv2.convexHull(contour)
drawing = np.zeros(img.shape, np.uint8)
cv2.drawContours(drawing, [contour], 0, (0, 255, 0), 2)
cv2.drawContours(drawing, [hull], 0, (0, 0, 255), 3)
imagepoints = self.handDetect.detectFingers(
contour, drawing)
if imagepoints.__len__() > 0:
self.calibrate.calibrate(imagepoints, drawing, frame)
glP, glM = self.calibrate.getGLPM()
# if you want to see calibration points, comment those lines
# draw objects on hand with opengl
self.demo = DemoAR(halfframe, glP, glM)
self.demo.main()
cv2.imshow('output', drawing)
# Keyboard OP
k = cv2.waitKey(10)
if k == 27: # press ESC to exit
exit(0)
elif k == ord('b'): # press 'b' to capture the background
self.handDetect.bgModel = cv2.BackgroundSubtractorMOG2(
0, self.bgSubThreshold)
self.isBgCaptured = 1
print('Background Captured')
elif k == ord('r'): # press 'r' to reset the background
self.handDetect.bgModel = None
self.isBgCaptured = 0
self.handDetect.setDoStart(False)
print('Reset BackGround')
elif k == ord('s'):
self.handDetect.setDoStart(True)
print('Start Calibration')
bd=0,
highlightthickness=0)
canvas.pack()
tk.update()
# Create ball
ball = Ball(canvas, 'red', eng)
i = 0
count = 1000
while i < count:
ball.draw_experiment(math.floor(i / 200))
tk.update_idletasks()
tk.update()
time.sleep(0.01)
i = i + 1
ball.gazeSubscriber.stop()
tk.destroy()
mean_error = eng.ExperimentData()
print('Mean Error: ' + str(mean_error) + 'px')
if __name__ == '__main__':
print("Start MatLab Engine:\n ")
eng = matlab.engine.start_matlab()
print("Start Calibrate:\n ")
Calibrate.calibrate(eng)
print("Start experiment:\n")
Experiment.experiment(eng)
