def __init__(self):
super(Webcamera, self).__init__()
self.history = []
self.colors = ImageColors()
self.filter = FilterHelper()
self.OnCapture = event.Event()
self.OnCaptureComplete = event.Event()
class Webcamera(Webcam):
#TODO calculate the right threshold
__daylightThreshold = 75
def __init__(self):
super(Webcamera, self).__init__()
self.history = []
self.colors = ImageColors()
self.filter = FilterHelper()
self.OnCapture = event.Event()
self.OnCaptureComplete = event.Event()
def capture(self, duration=10, delay=0.01, persist=False):
try:
self.history = []
_start = time()
_running = True
while _running:
args = []
_filename = str(uuid.uuid4()) + '.jpg'
_image = self.save_image(_filename, persist)
_intensity = self.image_average_intensity(_image)
_daylight = self.daytime(_intensity)
_mcc = self.image_most_common_colour(_image)
_size = (_image.height * _image.width)
_image_data = WebImage(_filename, _intensity, _daylight, _mcc, _size, localtime(), delay)
self.history.append(_image_data)
_elapsed = (time() - _start)
args.append(_elapsed)
args.append(localtime())
args.append(_image_data)
self.OnCapture(self, args)
if _elapsed >= duration: _running = False
#TODO add save and read back capability
self.OnCaptureComplete(self, None)
except Exception as ex:
print(ex)
def save_image(self, filename=None, persist=True):
try:
if filename == None: filename = str(uuid.uuid4()) + '.jpg'
_image = urlopen('{0}/axis-cgi/jpg/image.cgi'.format(self.url)).read()
# convert directly to an Image instead of saving / reopening
# thanks to SO: http://stackoverflow.com/a/12020860/377366
_image = Image.open(io.BytesIO(_image))
if persist:
_image.save(filename)
return _image
except Exception as ex:
print(ex)
def image_average_intensity(self, image):
try:
pixels = numpy.array(image.getdata())
return numpy.mean(pixels)
except Exception as ex:
print(ex)
def image_most_common_colour(self, image):
try:
pixels = image.getdata()
colours = Counter(pixels).most_common(1) #Handy Python function for counting occurences.
rgb = colours[0][0]
frequency = colours[0][1]
name = get_closest_colour(rgb)
return (rgb, frequency, name, (frequency/len(pixels)))
except Exception as ex:
print(ex)
def daytime(self, intensity=0):
return intensity >= self.__daylightThreshold
def plot_history_intensity(self):
try:
caption = 'average intensity over time'
with plt.xkcd():
fig, ax = plt.subplots(nrows=2, ncols=1, sharex=False, sharey=False, tight_layout=True, figsize=(9, 4.5))
fig.suptitle(caption, fontsize=18, fontweight='bold')
x = [datetime.fromtimestamp(mktime(wimage.time)) for wimage in self.history]
y = [wimage.intensity for wimage in self.history]
#ax.set_xticklabels(x, fontsize='small')
ax[0].plot(x, y)
y1 = self.filter.filterData(y, 5)
ax[1].plot(x, y1)
plt.show()
except Exception as ex:
print(ex)
def __image_rotate(self, image, angle=180):
(h, w) = image.shape[:2]
image_center = (w / 2, h / 2)
# rotate the image by 180 degrees
rotation_matrix = cv2.getRotationMatrix2D(image_center, angle, 1.0)
rotated = cv2.warpAffine(image, rotation_matrix, (w, h))
return rotated
def __preprocess_image(self, image, ratio=0.75, blur=21):
#resize to make things a bit faster, convert to grayscale,
# and apply some gaussian blur to reduce aliasing and pixel differences
_image = image.resize((int(image.width*ratio), int(image.height*ratio)))
_image = cv2.GaussianBlur(_image, (blur, blur), 0)
return _image
def funkyfy(self, image=None, colorrange=([0,100,0],[255, 255,120]), useopencv=True):
#Not the most elegant solution
try:
if colorrange ==None:
colorrange = ([numpy.random.randint(0,255), numpy.random.randint(0,255), numpy.random.randint(0,255)],
[numpy.random.randint(0,255), numpy.random.randint(0,255), numpy.random.randint(0,255)])
if image == None:
image = self.save_image(persist=False) #gets a webcam image
image = image.resize((int(image.width*0.5), int(image.height*0.5)))
image_bgr = cv2.cvtColor(numpy.array(image), cv2.COLOR_RGB2BGR)
image_bgr = self.__image_rotate(image_bgr, 2)
lower = numpy.array(colorrange[0], dtype="uint8")
upper = numpy.array(colorrange[1], dtype="uint8")
mask = cv2.inRange(image_bgr, lower, upper)
output = cv2.bitwise_and(image_bgr, image_bgr, mask = mask)
b, g, r = cv2.split(output)
g[g>0] = 255
output = cv2.merge((g,b,r))
for row in range(130, len(output)):
for col in range(len(output[row])):
tup = (output[row][col])
if any(v != 0 for v in tup):
image_bgr[row][col] = tup
image_bgr = self.__image_rotate(image_bgr, -2)
return image_bgr, output
except Exception as ex:
print(ex)
#TODO finetune mincontour,
# Gets two images and calculates the difference in pixels
def detect_motion(self, interval=0.5):
min_contour_area = 5
max_contour_area = 1250
retval = False
threshold = 65
try:
#TODO write code to change static image if night
_image_static = None
_image_static = self.save_image(persist=False)
_image_static = cv2.cvtColor(numpy.array(_image_static), cv2.COLOR_RGB2GRAY)
_image_static = cv2.GaussianBlur(_image_static, (21, 21), 0)
sleep(interval) #defaults to one second
_image_dynamic = self.save_image(persist=False)
_image_dynamic = cv2.cvtColor(numpy.array(_image_dynamic), cv2.COLOR_RGB2GRAY)
_image_dynamic1 = cv2.GaussianBlur(_image_dynamic, (21, 21), 0)
# ideas from http://docs.opencv.org/master/d4/d73/tutorial_py_contours_begin.html#gsc.tab=0
_delta = cv2.absdiff(_image_dynamic1, _image_static)
_threshold = cv2.threshold(_delta, 25, 255, cv2.THRESH_BINARY)[1]
# dilate the thresholded image to fill in holes, then find contour on thresholded image
_threshold = cv2.dilate(_threshold, None, iterations=2)
#TODO evaluate different parameters
(img, contours, _) = cv2.findContours(_threshold.copy(), cv2.RETR_EXTERNAL, cv2.CHAIN_APPROX_SIMPLE)
#(img, contours, _) = cv2.findContours(_threshold.copy(), cv2.RETR_TREE, cv2.CHAIN_APPROX_SIMPLE)
# loop over the contours
for contour in contours:
# if the contour is too small, ignore it
_area = cv2.contourArea(contour)
if _area < min_contour_area or _area > max_contour_area:
continue # skip to the next
# compute the bounding box for the contour, draw it on the frame,
retval = True
(x, y, w, h) = cv2.boundingRect(contour)
cv2.rectangle(_image_dynamic, (x, y), (x + w, y + h), (0, 12, 255), 2)
#cv2.ellipse(_image_dynamic, (x, y+20), (10, 20), 90, 0, 360, (255, 0, 0), 2)
#cv2.ellipse(_image_dynamic1,(200,200),(80,50),45,0,360,(0,0,255),1)
# draw the text and timestamp on the frame
#cv2.putText(_image_dynamic1, "Motion: {}".format(text), (10, 20), cv2.FONT_HERSHEY_SIMPLEX, 0.5, (0, 0, 255), 2)
#cv2.putText(_image_dynamic1, datetime.now().strftime("%A %d %B %Y %I:%M:%S%p"), (10, _image_dynamic1.shape[0] - 10), cv2.FONT_HERSHEY_SIMPLEX, 0.35, (0, 0, 255), 1)
return retval, _image_dynamic
except Exception as ex:
print(ex)
def detect_motion_new(self, winName, interval=1):
#Implemented after talking with Dr Smart about image averaging and background extraction
min_contour_area = 25
max_contour_area = 1250
retval = False
threshold = 65
try:
_image_static = None
_image_static = self.save_image(persist=False)
_image_static = cv2.cvtColor(numpy.array(_image_static), cv2.COLOR_RGB2GRAY)
_image_static = cv2.GaussianBlur(_image_static, (21, 21), 0)
accumulator = numpy.float32(_image_static)
while True:
sleep(interval)
_image_static = self.save_image(persist=False)
_image_static = cv2.cvtColor(numpy.array(_image_static), cv2.COLOR_RGB2GRAY)
_image_static = cv2.GaussianBlur(_image_static, (21, 21), 0)
cv2.accumulateWeighted(numpy.float32(_image_static), accumulator, 0.1)
_image_static = cv2.convertScaleAbs(accumulator)
_image_dynamic = self.save_image(persist=False)
_image_dynamic1 = cv2.cvtColor(numpy.array(_image_dynamic), cv2.COLOR_RGB2GRAY)
_image_dynamic1 = cv2.GaussianBlur(_image_dynamic1, (21, 21), 0)
# ideas from http://docs.opencv.org/master/d4/d73/tutorial_py_contours_begin.html#gsc.tab=0
_delta = cv2.absdiff(_image_dynamic1, _image_static)
_threshold = cv2.threshold(_delta, 17, 255, cv2.THRESH_BINARY)[1]
# dilate the thresholded image to fill in holes, then find contour on thresholded image
_threshold = cv2.dilate(_threshold, None, iterations=5)
(img, contours, _) = cv2.findContours(_threshold.copy(), cv2.RETR_EXTERNAL, cv2.CHAIN_APPROX_SIMPLE)
dyn = cv2.cvtColor(numpy.array(_image_dynamic), cv2.COLOR_RGB2GRAY)
# loop over the contours
for contour in contours:
# if the contour is too small, ignore it
_area = cv2.contourArea(contour)
if _area < min_contour_area: # or _area > max_contour_area:
continue # skip to the next
# compute the bounding box for the contour, draw it on the frame,
(x, y, w, h) = cv2.boundingRect(contour)
#cv2.rectangle(dyn, (x, y), (x + w, y + h), (0, 12, 255), 2)
cv2.ellipse(dyn, (x+5, y+25), (10, 20), 90, 0, 360, (255, 0, 0), 2)
cv2.imshow(winName, numpy.hstack([dyn, _threshold]))
key = cv2.waitKey(10)
if key == 27:
cv2.destroyWindow(winName)
break
except Exception as ex:
print(ex)
#TODO detect Event
def detect_event(self):
crop_coord = (180, 320, 555, 490)
min_contour_area = 1000
max_contour_area = 3000
event_count = 0
threshold = 65
try:
#convert to grayscale,
# and apply some gaussian blur to reduce aliasing (see wikipedia)
_image_dynamic = Image.open('event_day.jpg').crop(crop_coord) #self.save_image(persist=False).crop(crop_coord)
_image_dynamic = self.__preprocess_image(_image_dynamic, 1)
_image_static = Image.open('nopeeps_day.jpg').crop(crop_coord)
_image_static = self.__preprocess_image(_image_static, 1)
# ideas from http://docs.opencv.org/master/d4/d73/tutorial_py_contours_begin.html#gsc.tab=0
_delta = cv2.absdiff(_image_dynamic, _image_static)
#cv2.imshow('delta', _delta)
_threshold = cv2.threshold(_delta, 25, 255, cv2.THRESH_BINARY)[1]
#cv2.imshow('thres', _threshold)
# dilate the thresholded image to fill in holes, then find contour on thresholded image
_threshold = cv2.dilate(_threshold, None, iterations=2)
(img, contours, _) = cv2.findContours(_threshold.copy(), cv2.RETR_EXTERNAL, cv2.CHAIN_APPROX_SIMPLE)
# loop over the contours
text = 'None'
for contour in contours:
# if the contour is too small, ignore it
_area = cv2.contourArea(contour)
if _area < min_contour_area or _area > max_contour_area:
continue # skip to the next
# compute the bounding box for the contour, draw it on the frame,
event_count +=1
(x, y, w, h) = cv2.boundingRect(contour)
cv2.rectangle(_image_dynamic, (x, y), (x + w, y + h), (0, 255, 0), 2)
text = "Detected"
# draw the text and timestamp on the frame
return (event_count > 0), event_count, _image_dynamic
except Exception as ex:
print(ex)
