def stream(self):
Handler.stream(self)
print('[INFO] Overriding stream method...')
# Initialise capture
capture = _capture(src=application.CAPTURING_DEVICE,
use_pi_camera=application.USE_PI_CAMERA,
resolution=application.RESOLUTION,
frame_rate=application.FRAME_RATE)
if application.USE_PI_CAMERA:
print('[INFO] Warming up pi camera...')
else:
print('[INFO] Warming up camera...')
time.sleep(2.0)
print('[INFO] Start capturing...')
while True:
# Read a frame from capture
frame = capture.read()
# Down size frame to 50% (to increase performance on Raspberry Pi)
frame = _frame.scale(frame=frame, scale=0.5)
# Convert frame to gray (to increase performance on Raspberry Pi)
gray = cv2.cvtColor(frame, cv2.COLOR_BGR2GRAY)
# Dlib detection (set `up_sample=0` to increase performance on Raspberry Pi)
detections = _dlib.frontal_face_detector(frame=gray, up_sample=0)
# Up size frame to 50% (how the frame was before down sizing)
frame = _frame.scale(frame=frame, scale=2)
# If Dlib returns any detection
for det in detections:
# Up size coordinate to 50% (according to the frame size before down sizing)
coordinates = {
'left': det.left() * 2,
'top': det.top() * 2,
'right': det.right() * 2,
'bottom': det.bottom() * 2
}
# Draw box around detection with text on the frame
frame = _draw.rectangle(frame=frame,
coordinates=coordinates,
text='Detected')
# Determine the facial landmarks for the face region
# (using gray image to increase performance on Raspberry Pi)
landmarks = _dlib.shape_predictor_68_face_landmarks(gray, det)
# Draw circle around landmarks on the frame
for (x, y) in landmarks:
frame = _draw.circle(frame=frame,
coordinates={
'left': x * 2,
'top': y * 2
},
radius=0.5)
# Write date time on the frame
frame = _draw.text(frame=frame,
coordinates={
'left': application.WIDTH - 150,
'top': application.HEIGHT - 20
},
text=time.strftime('%d/%m/%Y %H:%M:%S',
time.localtime()),
font_color=(0, 0, 255))
# Convert frame into buffer for streaming
retval, buffer = cv2.imencode('.jpg', frame)
# Write buffer to HTML Handler
self.wfile.write(b'--FRAME\r\n')
self.send_header('Content-Type', 'image/jpeg')
self.send_header('Content-Length', len(buffer))
self.end_headers()
self.wfile.write(buffer)
self.wfile.write(b'\r\n')
def stream(self):
Handler.stream(self)
print('[INFO] Overriding stream method...')
# Initialise capture
capture = _capture(src=application.CAPTURING_DEVICE,
use_pi_camera=application.USE_PI_CAMERA,
resolution=application.RESOLUTION,
frame_rate=application.FRAME_RATE)
if application.USE_PI_CAMERA:
print('[INFO] Warming up pi camera...')
else:
print('[INFO] Warming up camera...')
time.sleep(2.0)
print('[INFO] Start capturing...')
while True:
# Read a frame from capture
frame = capture.read()
# Down size frame to 50% (to increase performance on Raspberry Pi)
frame = _frame.scale(frame=frame, scale=0.5)
# Convert frame to gray (to increase performance on Raspberry Pi)
gray = cv2.cvtColor(frame, cv2.COLOR_BGR2GRAY)
# OpenCV detection
detections = _opencv.haarcascade_frontalface_default_detector(
gray, scale_factor=1.0, min_neighbours=5, min_size=(15, 15))
# Up size frame to 50% (how the frame was before down sizing)
frame = _frame.scale(frame=frame, scale=2)
# If returns any detection
for (x, y, w, h) in detections:
# Up size coordinate to 50% (according to the frame size before down sizing)
coordinates = {
'left': x * 2,
'top': y * 2,
'right': x * 2 + w * 2,
'bottom': y * 2 + h * 2
}
# Draw box around detection with text on the frame
frame = _draw.rectangle(frame=frame,
coordinates=coordinates,
text='Detected')
# Write date time on the frame
frame = _draw.text(frame=frame,
coordinates={
'left': application.WIDTH - 150,
'top': application.HEIGHT - 20
},
text=time.strftime('%d/%m/%Y %H:%M:%S',
time.localtime()),
font_color=(0, 0, 255))
# Convert frame into buffer for streaming
retval, buffer = cv2.imencode('.jpg', frame)
# Write buffer to HTML Handler
self.wfile.write(b'--FRAME\r\n')
self.send_header('Content-Type', 'image/jpeg')
self.send_header('Content-Length', len(buffer))
self.end_headers()
self.wfile.write(buffer)
self.wfile.write(b'\r\n')