def dir_thresh(image, thresh=(0, np.pi/2)):
gray = colorModel.rgbToYuv(image, gray.shape)
gray = colorModel.yuvToGrayscaleRgb(gray, gray.shape)
gaussianImg = img.copy()
gaussianDeviation = 1.0 ##############################################
gaussianFilterSize = math.floor(gaussianDeviation*3.0)
filters.gaussianBlur('RGB', 0, gaussianImg.load(),
gaussianImg.size, (gaussianFilterSize, gaussianFilterSize))
edgeDetection.canny('RGB', 0, img.load(), img.size,
gaussianImg.load(), 1, thresh)
dir_threshold = np.zeros_like(gray)
dir_threshold[(gray > thresh[0]) & (gray < thresh[1])] = 1
return dir_threshold
def mag_thresh(image, thresh=(1, 5)):
gray = colorModel.rgbToYuv(image, gray.shape)
gray = colorModel.yuvToGrayscaleRgb(gray, gray.shape)
gaussianImg = img.copy()
gaussianDeviation = 1.0 ##############################################
gaussianFilterSize = math.floor(gaussianDeviation*3.0)
filters.gaussianBlur('RGB', 0, gaussianImg.load(),
gaussianImg.size, (gaussianFilterSize, gaussianFilterSize))
edgeDetection.canny('RGB', 0, img.load(), img.size,
gaussianImg.load(), 1, thresh)
scaled_magnitude = np.uint8(255*gray_magnitude/np.max(gray_magnitude))
mag_thresh = np.zeros_like(scaled_magnitude)
mag_thresh[(scaled_magnitude > thresh[0]) & (scaled_magnitude < thresh[1])] = 1
return mag_thresh
def segLaplacian(self, colorModelTag, currentImageChannelIndex,
isOriginalImage, amplifier, threshold, maskMode):
"""
Laplacian
"""
img = self.imageService.openImage(isOriginalImage)
if img is None:
return
methodTimer = time.time()
if colorModelTag == 'RGB':
gaussianImg = img.copy()
gaussianDeviation = 1.0 ##############################################
gaussianFilterSize = math.floor(gaussianDeviation * 3.0)
filters.gaussianBlur(colorModelTag, currentImageChannelIndex,
gaussianImg.load(), gaussianImg.size,
(gaussianFilterSize, gaussianFilterSize))
edgeDetection.canny(colorModelTag,
currentImageChannelIndex,
img.load(),
img.size,
gaussianImg.load(),
amplifier,
threshold,
maskMode=maskMode)
methodTimer = time.time() - methodTimer
self.histogramService.saveHistogram(img=img, model=colorModelTag)
if colorModelTag == 'YUV':
colorModel.rgbToYuv(img.load(), img.size)
gaussianImg = img.copy()
gaussianDeviation = 1.0 ##############################################
gaussianFilterSize = math.floor(gaussianDeviation * 3.0)
filters.gaussianBlur(colorModelTag, currentImageChannelIndex,
gaussianImg.load(), gaussianImg.size,
(gaussianFilterSize, gaussianFilterSize))
edgeDetection.canny(colorModelTag,
currentImageChannelIndex,
img.load(),
img.size,
gaussianImg.load(),
amplifier,
threshold,
maskMode=maskMode)
methodTimer = time.time() - methodTimer
self.histogramService.saveHistogram(img=img, model=colorModelTag)
timerTemp = time.time()
colorModel.yuvToRgb(img.load(), img.size)
methodTimer = time.time() - methodTimer
if colorModelTag == 'HSL':
data = numpy.asarray(img, dtype="float")
data = colorModel.rgbToHsl(data)
gaussianData = numpy.copy(data)
gaussianDeviation = 1.0 ##############################################
gaussianFilterSize = math.floor(gaussianDeviation * 3.0)
filters.gaussianBlur(colorModelTag, currentImageChannelIndex,
gaussianData, gaussianData.shape,
(gaussianFilterSize, gaussianFilterSize))
edgeDetection.canny(colorModelTag,
currentImageChannelIndex,
gaussianData,
gaussianData.shape,
gaussianData,
amplifier,
threshold,
maskMode=maskMode)
methodTimer = time.time() - methodTimer
# data = numpy.copy(dataTemp)
self.histogramService.saveHistogram(data=data, model=colorModelTag)
timerTemp = time.time()
data = colorModel.hslToRgb(data)
img = Image.fromarray(
numpy.asarray(numpy.clip(data, 0, 255), dtype="uint8"))
methodTimer = time.time() - timerTemp + methodTimer
logFile = '{}/temp/log/segLaplacian.log'.format(self.appDir)
with open(logFile, "a+") as text_file:
text_file.write("Timer: {}: {}\n".format(colorModelTag,
methodTimer))
img.save('{}/temp/processingImage.png'.format(self.appDir))
imageComparison.calculateImageDifference(colorModelTag, logFile)
def detect(self, frame):
#img = cv2.cvtColor(frame, cv2.COLOR_BGR2GRAY)
colorModel.rgbToYuv(frame, frame.shape)
colorModel.yuvToGrayscaleRgb(frame, frame.shape)
roiy_end = frame.shape[0]
roix_end = frame.shape[1]
img = Image.fromarray(
numpy.asarray(numpy.clip(frame, 0, 255), dtype="uint8"))
img.show()
roi = frame[self.road_horizon:roiy_end, 0:roix_end]
print(roi)
#blur = cv2.medianBlur(roi, 5)
blur = numpy.copy(frame)
filters.medianFilter('RGB', 0, blur, blur.shape, (5, 5))
# contours = cv2.Canny(blur, 60, 120)
contours = numpy.copy(blur)
gaussianData = numpy.copy(blur)
gaussianDeviation = 1.0 ##############################################
gaussianFilterSize = math.floor(gaussianDeviation * 3.0)
filters.gaussianBlur('RGB', 0, gaussianData, gaussianData.shape,
(gaussianFilterSize, gaussianFilterSize))
edgeDetection.canny('RGB',
0,
contours,
contours.shape,
gaussianData,
amplifier=1.0,
threshold=(60, 120))
# if self.prob_hough:
# lines = cv2.HoughLinesP(contours, 1, numpy.pi/180, self.vote, minLineLength=30, maxLineGap=100)
# else:
lines = self._standard_hough(contours, self.vote)
if lines is not None:
# find nearest lines to center
lines = lines + numpy.array([
0, self.road_horizon, 0, self.road_horizon
]).reshape(
(1, 1, 4)
) # scale points from ROI coordinates to full frame coordinates
left_bound = None
right_bound = None
for l in lines:
# find the rightmost line of the left half of the frame and the leftmost line of the right half
for x1, y1, x2, y2 in l:
theta = numpy.abs(numpy.arctan2(
(y2 - y1), (x2 - x1))) # line angle WRT horizon
if theta > self.roi_theta: # ignore lines with a small angle WRT horizon
dist = self._base_distance(x1, y1, x2, y2,
frame.shape[1])
if left_bound is None and dist < 0:
left_bound = (x1, y1, x2, y2)
left_dist = dist
elif right_bound is None and dist > 0:
right_bound = (x1, y1, x2, y2)
right_dist = dist
elif left_bound is not None and 0 > dist > left_dist:
left_bound = (x1, y1, x2, y2)
left_dist = dist
elif right_bound is not None and 0 < dist < right_dist:
right_bound = (x1, y1, x2, y2)
right_dist = dist
if left_bound is not None:
left_bound = self._scale_line(left_bound[0], left_bound[1],
left_bound[2], left_bound[3],
frame.shape[0])
if right_bound is not None:
right_bound = self._scale_line(right_bound[0], right_bound[1],
right_bound[2], right_bound[3],
frame.shape[0])
return [left_bound, right_bound]