def main():
log = logging.getLogger("main")
# creating exit mask from points, where we will be counting our vehicles
base = np.zeros(SHAPE + (3, ), dtype='uint8')
exit_mask = cv2.fillPoly(base, EXIT_PTS, (255, 255, 255))[:, :, 0]
# there is also bgslibrary, that seems to give better BG substruction, but
# not tested it yet
bg_subtractor = cv2.createBackgroundSubtractorMOG2(history=500,
detectShadows=True)
# processing pipline for programming conviniance
pipeline = PipelineRunner(
pipeline=[
ContourDetection(bg_subtractor=bg_subtractor,
save_image=True,
image_dir=IMAGE_DIR),
# we use y_weight == 2.0 because traffic are moving vertically on video
# use x_weight == 2.0 for horizontal.
VehicleCounter(exit_masks=[exit_mask], y_weight=2.0),
Visualizer(image_dir=IMAGE_DIR),
CsvWriter(path='./', name='report.csv')
],
log_level=logging.DEBUG)
# Set up image source
# You can use also CV2, for some reason it not working for me
cap = skvideo.io.vreader(VIDEO_SOURCE)
# skipping 500 frames to train bg subtractor
train_bg_subtractor(bg_subtractor, cap, num=500)
_frame_number = -1
frame_number = -1
for frame in cap:
if not frame.any():
log.error("Frame capture failed, stopping...")
break
# real frame number
_frame_number += 1
# skip every 2nd frame to speed up processing
if _frame_number % 2 != 0:
continue
# frame number that will be passed to pipline
# this needed to make video from cutted frames
frame_number += 1
# plt.imshow(frame)
# plt.show()
# return
pipeline.set_context({
'frame': frame,
'frame_number': frame_number,
})
pipeline.run()
def main():
log = logging.getLogger("main")
# creating exit mask from points, where we will be counting our vehicles
base = np.zeros(SHAPE + (3, ), dtype='uint8')
exit_mask = cv2.fillPoly(base, EXIT_PTS, (255, 255, 255))[:, :, 0]
# there is also bgslibrary, that seems to give better BG substruction, but
# not tested it yet
bg_subtractor = cv2.createBackgroundSubtractorMOG2(history=500,
detectShadows=True)
# processing pipeline for programming conviniance
pipeline = PipelineRunner(
pipeline=[
ContourDetection(bg_subtractor=bg_subtractor,
save_image=True,
image_dir=IMAGE_DIR),
# we use y_weight == 2.0 because traffic are moving vertically on video
# use x_weight == 2.0 for horizontal.
VehicleCounter(exit_masks=[exit_mask], y_weight=2.0, x_weight=1.5),
Visualizer(image_dir=IMAGE_DIR),
CsvWriter(path='./', name=REPORT_NAME)
],
log_level=logging.DEBUG)
# Set up image source
# You can use also CV2, for some reason it not working for me
cap = cv2.VideoCapture(VIDEO_SOURCE)
# skipping 100 frames to train bg subtractor
train_bg_subtractor(bg_subtractor, cap, num=500)
frame_number = -1
while True:
ret, frame = cap.read()
if frame is None:
log.error("Frame capture failed, stopping...")
break
frame_number += 1
# plt.imshow(frame)
# plt.show()
# return
pipeline.set_context({
'frame': frame,
'frame_number': frame_number,
})
pipeline.run()
def main():
log = logging.getLogger("main")
# creating exit mask from points, where we will be counting our vehicles
base = np.zeros(SHAPE + (3, ), dtype='uint8')
exit_mask = cv2.fillPoly(base, EXIT_PTS, (255, 255, 255))[:, :, 0]
# there is also bgslibrary, that seems to give better BG substruction, but
# not tested it yet
bg_subtractor = cv2.createBackgroundSubtractorMOG2(history=500,
detectShadows=True)
# processing pipline for programming conviniance
pipeline = PipelineRunner(
pipeline=[
ContourDetection(bg_subtractor=bg_subtractor,
min_contour_width=MIN_COUNTOUR_WIDTH,
min_contour_height=MIN_COUNTOUR_HEIGHT,
save_image=True,
image_dir=IMAGE_DIR),
# we use y_weight == 2.0 because traffic are moving vertically on video
# use x_weight == 2.0 for horizontal.
VehicleCounter(exit_masks=[exit_mask], x_weight=2.0),
Visualizer(image_dir=IMAGE_DIR),
# CsvWriter(path='./', name='report.csv')
],
log_level=logging.DEBUG)
# Set up image source
# You can use also CV2, for some reason it not working for me
# cap = skvideo.io.vreader(VIDEO_SOURCE)
# skipping 500 frames to train bg subtractor
if VIDEO_SOURCE:
cap = skvideo.io.vreader(VIDEO_SOURCE)
train_bg_subtractor_video(bg_subtractor, cap, num=500)
cap = skvideo.io.vreader(VIDEO_SOURCE)
run_video(pipeline, cap)
else:
camera = PiCamera()
camera.resolution = (640, 368)
camera.framerate = 32
rawCapture = PiRGBArray(camera, size=(640, 368))
# allow the camera to warmup
time.sleep(0.1)
train_bg_subtractor_pi(bg_subtractor, camera, rawCapture, num=500)
run_pi(pipeline, camera, rawCapture)
cap.release()
cv2.destroyAllWindows()
def main():
log = logging.getLogger("main")
base = np.zeros(SHAPE + (3, ), dtype='uint8')
exit_mask = cv2.fillPoly(base, EXIT_PTS, (255, 255, 255))[:, :, 0]
bg_subtractor = cv2.createBackgroundSubtractorMOG2(history=500,
detectShadows=True)
pipeline = PipelineRunner(
pipeline=[
ContourDetection(bg_subtractor=bg_subtractor,
save_image=True,
image_dir=IMAGE_DIR),
# y_weight == 2.0 vertical
# x_weight == 2.0 for horizontal. in below line
VehicleCounter(exit_masks=[exit_mask], y_weight=2.0),
Visualizer(image_dir=IMAGE_DIR),
CsvWriter(path='./', name='report.csv')
],
log_level=logging.DEBUG)
cap = cv2.VideoCapture(VIDEO_SOURCE)
train_bg_subtractor(bg_subtractor, cap, num=500)
_frame_number = -1
frame_number = -1
while cap.isOpened():
ret, frame = cap.read()
_frame_number += 1
if _frame_number % 2 != 0:
continue
frame_number += 1
# plt.imshow(frame)
# plt.show()
# return
pipeline.set_context({
'frame': frame,
'frame_number': frame_number,
})
pipeline.run()
print("End of Video Reached......")
def main():
log = logging.getLogger("main")
base = np.zeros(SHAPE + (3, ), dtype='uint8')
exit_mask = cv2.fillPoly(base, EXIT_PTS, (255, 255, 255))[:, :, 0]
bg_subtractor = cv2.createBackgroundSubtractorMOG2(history=500,
detectShadows=True)
pipeline = PipelineRunner(pipeline=[
ContourDetection(bg_subtractor=bg_subtractor,
save_image=True,
image_dir=IMAGE_DIR),
VehicleCounter(exit_masks=[exit_mask], y_weight=2.0),
Visualizer(image_dir=IMAGE_DIR),
CsvWriter(path='./', name='report.csv')
],
log_level=logging.DEBUG)
cap = skvideo.io.vreader(VIDEO_SOURCE)
train_bg_subtractor(bg_subtractor, cap, num=500)
_frame_number = -1
frame_number = -1
for frame in cap:
if not frame.any():
log.error("Frame capture failed, stopping...")
break
_frame_number += 1
if _frame_number % 2 != 0:
continue
frame_number += 1
# plt.imshow(frame)
# plt.show()
# return
pipeline.set_context({
'frame': frame,
'frame_number': frame_number,
})
pipeline.run()
def main():
skvideo.setFFmpegPath(
"D:\\Computer Vision\\Homework1\\opencv_traffic_counting\\ffmpeg-20190519-fbdb3aa-win64-static\\bin"
)
log = logging.getLogger("main")
base = np.zeros(SHAPE + (3, ), dtype='uint8')
exit_mask = cv2.fillPoly(base, EXIT_PTS, (255, 255, 255))[:, :, 0]
bg_subtractor = cv2.createBackgroundSubtractorMOG2(history=500,
detectShadows=True)
pipeline = ProcessPipelineRunner(pipeline=[
ContourDetection(bg_subtractor=bg_subtractor,
save_image=True,
image_dir=Image_dir),
VehicleCounter(exit_masks=[exit_mask], y_weight=2.0),
Vis(image_dir=Image_dir),
write_csv(path='./', name='report.csv')
],
log_level=logging.DEBUG)
cap = skvideo.io.vreader(Vid_src)
# skip num frames and train the background subtractor it will identify the background for all the frames
bg_Subtractor_train(bg_subtractor, cap, num=1000)
_frame_number = -1
frame_number = -1
for frame in cap:
if not frame.any():
log.error("Frame capture failed, stopping...")
break
_frame_number += 1
if _frame_number % 2 != 0:
continue
frame_number += 1
pipeline.set_context({
'frame': frame,
'frame_number': frame_number,
})
pipeline.run()
def main():
log = logging.getLogger("main")
# creating an exit mask from points where we will be counting out vehicles
base = np.zeros(Shape + (3, ), dtype='uint8')
exit_mask = cv2.fillPoly(base, Exit_pts, (255, 255, 255))[:, :, 0]
bg_subtractor = cv2.createBackgroundSubtractorMOG2(history=500,
detectShadows=True)
# processing pipeline for programming convenience
pipeline = PipelineRunner(
pipeline=[
ContourDetection(bg_subtractor=bg_subtractor,
save_image=True,
image_dir=Image_dir),
# we use y_weight == 2.0 because traffic are moving vertically on video
# use x_weight == 2.0 for horizontal
VehicleCounter(exit_masks=[exit_mask], y_weight=2.0),
Visualizer(image_dir=Image_dir),
CsvWriter(path='./', name='report.csv')
],
log_level=logging.DEBUG)
#setting up image source
cap = skvideo.io.vreader(Video_source)
# skip 500 frames to train bg subtractor
train_bg_subtractor(bg_subtractor, cap, num=500)
_frame_number = -1
frame_number = -1
for frame in cap:
if not frame.any():
log.error("frame capture failed, stopping...")
break
# real frame number
_frame_number += 1
#skipping every 2nd frame to speed up processing
if _frame_number % 2 != 0:
continue
frame_number += 1
pipeline.set_context({'frame': frame, 'frame_number': frame_number})
pipeline.run()
def main():
# creating exit mask from points, where we will be counting our vehicles
base = np.zeros(SHAPE + (3, ), dtype='uint8')
exit_mask = cv2.fillPoly(base, EXIT_PTS, (255, 255, 255))[:, :, 0]
bg_subtractor = cv2.createBackgroundSubtractorMOG2(history=500,
detectShadows=True)
pipeline = PipelineRunner(pipeline=[
ContourDetection(
bg_subtractor=bg_subtractor, save_image=True, image_dir=IMAGE_DIR),
# we use y_weight == 2.0 because traffic are moving vertically on video
# use x_weight == 2.0 for horizontal.
VehicleCounter(exit_masks=[exit_mask], y_weight=2.0),
Visualizer(image_dir=IMAGE_DIR),
])
cap = skvideo.io.vreader(VIDEO_SOURCE)
train_bg_subtractor(bg_subtractor, cap, num=500)
_frame_number = -1
frame_number = -1
for frame in cap:
if not frame.any():
break
_frame_number += 1
if _frame_number % 2 != 0:
continue
frame_number += 1
pipeline.set_context({
'frame': frame,
'frame_number': frame_number,
})
pipeline.run()
exit_mask = cv2.fillPoly(base, EXIT_PTS, (255, 255, 255))[:, :, 0]
# there is also bgslibrary, that seems to give better BG substruction, but
# not tested it yet
bg_subtractor = cv2.createBackgroundSubtractorMOG2(history=500,
detectShadows=True)
# processing pipline for programming conviniance
pipeline = PipelineRunner(
pipeline=[
ContourDetection(bg_subtractor=bg_subtractor,
save_image=True,
image_dir=IMAGE_DIR),
# we use y_weight == 2.0 because traffic are moving vertically on video
# use x_weight == 2.0 for horizontal.
VehicleCounter(exit_masks=[exit_mask], y_weight=2.0),
Visualizer(image_dir=IMAGE_DIR),
CsvWriter(path='./', name='report.csv')
],
log_level=logging.DEBUG)
# skipping 500 frames to train bg subtractor
train_bg_subtractor(bg_subtractor, cap, num=500)
_frame_number = -1
frame_number = -1
cv2.namedWindow('op', cv2.WINDOW_NORMAL)
while True:
_, frame = cap.read()
# for frame in cap:
if not frame.any():
def main():
log = logging.getLogger("main")
# pick pixel distance start&end by double clicking
PIXEL_DISTANCE = select_pixel_distance(VIDEO_SOURCE)
if not PIXEL_DISTANCE:
print("No selection of PIXEL_DISTANCE!")
return
else:
print('PIXEL_DISTANCE: ')
print(PIXEL_DISTANCE)
PYHSICAL_DISTANCE = float(
input(
"Please enter the physical distance in meters of the pixel distance selected:"
))
METER_PER_PIXEL = PYHSICAL_DISTANCE / PIXEL_DISTANCE
# draw polygons using mouse to pick exit points
EXIT_PTS = select_exit_zones(VIDEO_SOURCE)
if not EXIT_PTS:
print("No selection of exit zone!")
return
else:
EXIT_PTS = np.array(EXIT_PTS)
print('EXIT_PTS: ')
print(EXIT_PTS)
# Set up image source
cap = cv2.VideoCapture(VIDEO_SOURCE)
width = int(cap.get(cv2.CAP_PROP_FRAME_WIDTH))
height = int(cap.get(cv2.CAP_PROP_FRAME_HEIGHT))
fps = int(cap.get(cv2.CAP_PROP_FPS))
n_frames = int(cap.get(cv2.CAP_PROP_FRAME_COUNT))
print(width, height)
fourcc = cv2.VideoWriter_fourcc(*'mp4v')
out = cv2.VideoWriter(VIDEO_OUT_DEST, fourcc, fps, (width, height))
# creating exit mask from points, where we will be counting our vehicles
base = np.zeros((height, width) + (3, ), dtype='uint8')
exit_mask = cv2.fillPoly(base, EXIT_PTS, (255, 255, 255))[:, :, 0]
# there is also bgslibrary, that seems to give better BG substruction, but
# not tested it yet
bg_subtractor = cv2.createBackgroundSubtractorMOG2(history=500,
detectShadows=True)
# processing pipline for programming conviniance
pipeline = PipelineRunner(
pipeline=[
ContourDetection(bg_subtractor=bg_subtractor,
min_contour_width=int(MIN_CONTOUR_RATIO * height),
min_contour_height=int(MIN_CONTOUR_RATIO *
height),
save_image=False,
image_dir=IMAGE_DIR),
# we use y_weight == 2.0 because traffic are moving vertically on video
# use x_weight == 2.0 for horizontal.
VehicleCounter(use_physical_speed=USE_PHYSICAL_SPEED,
meter_per_pixel=METER_PER_PIXEL,
fps=fps,
avg_speed_interval=AVG_SPEED_INTERVAL,
exit_masks=[exit_mask],
y_weight=2.0,
path_size=10),
Visualizer(use_physical_speed=USE_PHYSICAL_SPEED,
video_out=out,
image_dir=IMAGE_DIR,
save_image=False),
CsvWriter(path='./', name='report.csv')
],
log_level=logging.DEBUG)
# skipping 500 frames to train bg subtractor, close video and reopen
train_bg_subtractor(bg_subtractor, cap, num=500)
cap.release()
frame_number = -1
frame_time_sec = -1.0 / fps
cap = cv2.VideoCapture(VIDEO_SOURCE)
while (cap.isOpened()):
ret, frame = cap.read()
if ret == True:
# calculate the frame time in second
frame_time_sec += 1.0 / fps
# frame number that will be passed to pipline
# this needed to make video from cutted frames
frame_number += 1
# plt.imshow(frame)
# plt.show()
# return
pipeline.set_context({
'frame': frame,
'frame_number': frame_number,
'frame_time_sec': frame_time_sec
})
pipeline.run()
if cv2.waitKey(1) & 0xFF == ord('q'):
break
else:
break
# Release everything if job is finished
cap.release()
out.release()
cv2.destroyAllWindows()