def contours_detector(frame, typeBinariz=0, binarizParams=(9, 2.5),
minSizeObj=(24, 24), maxSizeObj=(144, 144),
aspectRatioObjInterval=(0.6, 1.3), compactObjInterval=(0.01, 0.2)):
# cut left and right parts
fr_cut_main = cv.GaussianBlur(frame, (5, 5), 0)
fr_cut_left = frproc.cut_frame(fr_cut_main, (0, 0), (int(0.3 * fr_cut_main.shape[1]), fr_cut_main.shape[0]))
fr_cut_right = frproc.cut_frame(fr_cut_main, (int(0.7 * fr_cut_main.shape[1]), 0),
(fr_cut_main.shape[1], fr_cut_main.shape[0]))
# equalize histogram for left and right parts of frame
fr_cut_left = cv.createCLAHE(2.0, (5, 5)).apply(fr_cut_left)
fr_cut_left = cv.GaussianBlur(fr_cut_left, (5, 5), 0)
fr_cut_right = cv.createCLAHE(2.0, (5, 5)).apply(fr_cut_right)
fr_cut_right = cv.GaussianBlur(fr_cut_right, (5, 5), 0)
# binarization type
if typeBinariz == 0:
tr_left = cv.adaptiveThreshold(fr_cut_left, 255, cv.ADAPTIVE_THRESH_MEAN_C, cv.THRESH_BINARY_INV,
binarizParams[0], binarizParams[1])
tr_main = cv.adaptiveThreshold(fr_cut_main, 255, cv.ADAPTIVE_THRESH_MEAN_C, cv.THRESH_BINARY_INV,
binarizParams[0], binarizParams[1])
tr_right = cv.adaptiveThreshold(fr_cut_right, 255, cv.ADAPTIVE_THRESH_MEAN_C, cv.THRESH_BINARY_INV,
binarizParams[0], binarizParams[1])
elif typeBinariz == 1:
tr_left = cv.Canny(fr_cut_left, binarizParams[0], binarizParams[1])
tr_main = cv.Canny(fr_cut_main, binarizParams[0], binarizParams[1])
tr_right = cv.Canny(fr_cut_right, binarizParams[0], binarizParams[1])
elif typeBinariz == 2:
tr_left = imu.auto_canny(fr_cut_left)
tr_main = imu.auto_canny(fr_cut_main)
tr_right = imu.auto_canny(fr_cut_right)
else:
return
# concatenate parts in full frame
tr_main[0:fr_cut_main.shape[0], 0:(int(0.3 * fr_cut_main.shape[1]))] = tr_left[::]
tr_main[0:fr_cut_main.shape[0], int(0.7 * fr_cut_main.shape[1]):fr_cut_main.shape[1]] = tr_right[::]
cv.imshow("Test contours", tr_main)
_, contours, _ = cv.findContours(tr_main, cv.RETR_LIST, cv.CHAIN_APPROX_SIMPLE)
# sorting contours
contours = [c for c in contours
if maxSizeObj[0] >= cv.boundingRect(c)[2] >= minSizeObj[0] # width
and maxSizeObj[1] >= cv.boundingRect(c)[3] >= minSizeObj[1] # height
and (aspectRatioObjInterval[0] <=
(float(cv.boundingRect(c)[2]) / cv.boundingRect(c)[3]) <= aspectRatioObjInterval[1]) # aspect
and ((cv.arcLength(c, True) != 0) and
(compactObjInterval[0] <=
(float(cv.contourArea(c)) / (cv.arcLength(c, True) ** 2)) <= compactObjInterval[1]))] # compact
rects_signs = [cv.boundingRect(c) for c in contours]
return rects_signs
def verif():
with picamera.PiCamera() as camera:
camera.resolution = (
2592, 1944
) #défini la résolution de la caméra (max :(2592, 1944); min :(64,64))
camera.framerate = 15 #défini la fréuence d'image de la caméra (défini à 15 pour la résolution max)
camera.exposure_mode = 'auto' #mode d'exposition de la caméra (par défault : 'auto')
camera.start_preview(
) #ouverture de l'aperçu de la caméra (pas nécessaire dans notre cas mais pratiquepour le tests)
sleep(
2
) #ouverture de l'objectif pendant 5s (pour réglage luminosité ...)
camera.capture('/home/pi/varroasVerif.jpg'
) #capture de l'image et lien de stockage
camera.stop_preview() #fermeture de l'apercu
print('Vérification effectué')
sleep(2)
image = cv2.imread('varroasVerif.jpg')
gray = cv2.cvtColor(image,
cv2.COLOR_BGR2GRAY) #met l'image en noir et blanc
flougaussien = cv2.bilateralFilter(
gray, 6, 157, 157) #floûte l'image -> faire une moyenne des pixels
edge = imutils.auto_canny(flougaussien) #détermine les contours
(cnts, _) = cv2.findContours(
edge, cv2.RETR_EXTERNAL, cv2.CHAIN_APPROX_SIMPLE
) #ajoute dans une liste les contours qu'ils trouvent suivant les méthodes de recherche
compteurVerif = 0 #initialisation de la variable compteur
for c in cnts:
if (120 > cv2.contourArea(c) > 85
): #calcule le périmètre des contours trouvés (non fermé)
compteurVerif += 1
return compteurVerif
def preprocessing(self):
img1 = self.img[0:400, :]
_, img1 = cv2.threshold(img1, 25, 255, cv2.THRESH_TOZERO)
_, img1 = cv2.threshold(img1, 25, 255, cv2.THRESH_BINARY)
kernel = np.ones((9, 9), np.uint8)
img1 = cv2.morphologyEx(img1, cv2.MORPH_CLOSE, kernel)
img1 = cv2.erode(img1, (7, 7), iterations=1)
img1 = cv2.morphologyEx(img1, cv2.MORPH_OPEN, kernel)
img2 = self.img[400:500, :]
img2 = cv2.GaussianBlur(img2, (5, 5), np.sqrt(16))
_, img2 = cv2.threshold(img2, 1, 255, cv2.THRESH_BINARY)
img3 = self.img[500::, :]
kernel1 = np.ones((5, 5), np.uint8)
_, img3 = cv2.threshold(img3, 50, 255, cv2.THRESH_BINARY)
img3 = cv2.morphologyEx(img3, cv2.MORPH_CLOSE, kernel1)
img3 = cv2.erode(img3, (7, 7), iterations=3)
img3 = cv2.morphologyEx(img3, cv2.MORPH_OPEN, kernel1)
hold = np.zeros((self.img.shape[0], self.img.shape[1]), dtype=np.uint8)
hold[0:400, :] = img1
hold[400:500, :] = img2
hold[500::, :] = img3
hold = imu.auto_canny(hold)
hold = cv2.dilate(hold, (3, 3), iterations=3)
return hold
def scan(img):
# preprocess image
gray = cv2.cvtColor(img, cv2.COLOR_BGR2GRAY)
gray = cv2.bilateralFilter(gray, FILTER[0], FILTER[1], FILTER[2])
ret, gray = cv2.threshold(gray, THRESHOLD[0], THRESHOLD[1], 0)
edges = imutils.auto_canny(gray)
# extract contours
_, cnts, _ = cv2.findContours(edges.copy(), cv2.RETR_TREE,
cv2.CHAIN_APPROX_SIMPLE)
cnts = [c for c in cnts if cv2.contourArea(c) >= MIN_CARD_AREA]
card, c = None, None
if cnts:
# get largest contour
c = sorted(cnts, key=cv2.contourArea, reverse=True)[0]
# approximate the contour
peri = cv2.arcLength(c, True)
approx = cv2.approxPolyDP(c, 0.05 * peri, True)
pts = np.float32(approx)
x, y, w, h = cv2.boundingRect(c)
# Find center point of card by taking x and y average of the four corners.
# average = np.sum(pts, axis=0)/len(pts)
# cent_x = int(average[0][0])
# cent_y = int(average[0][1])
# center = [cent_x, cent_y]
# Warp card into 200x300 flattened image using perspective transform
card = util.flattener(img, pts, w, h)
card = util.cv2_to_pil(card).rotate(ROTATION)
return card, c, gray, edges
def read_test(image_path):
resp = urllib.request.urlopen(image_path)
image = np.asarray(bytearray(resp.read()), dtype="uint8")
image = cv2.imdecode(image, cv2.IMREAD_COLOR)
gray = cv2.cvtColor(image, cv2.COLOR_BGR2GRAY)
edged = imutils.auto_canny(gray)
'''
Find contours in the edge map, keeping only the largest one which
is presmumed to be the nail.
'''
cnts = cv2.findContours(edged.copy(), cv2.RETR_EXTERNAL,
cv2.CHAIN_APPROX_SIMPLE)
cnts = cnts[0] if imutils.is_cv2() else cnts[1]
c = max(cnts, key=cv2.contourArea)
# --- Extract the nail and resize it to a canonical width and height ---
(x, y, w, h) = cv2.boundingRect(c)
nail = gray[y:y + h, x:x + w]
nail = cv2.resize(nail, (200, 100))
fd = hog(nail,
orientations=9,
pixels_per_cell=(10, 10),
cells_per_block=(2, 2),
transform_sqrt=True,
block_norm="L1")
return fd.reshape(1, -1)
def process_markers(img):
hi, wi, _ = img.shape
blur = cv2.medianBlur(img, 31)
edge = imutils.auto_canny(blur)
edge = cv2.dilate(edge, np.ones((5, 5)), iterations=2)
_, contours, hierarchy = cv2.findContours(np.copy(edge), cv2.RETR_TREE, cv2.CHAIN_APPROX_SIMPLE)
img_tmp = np.copy(img)
cv2.drawContours(img_tmp, contours, -1, (100, 0, 0), 2)
cv2.drawContours(edge, contours, -1, 100, 2)
cnt = 0
roi_ptr = []
padding = 10
for i in range(len(hierarchy[0])):
if hierarchy[0][i][3] == -1:
x, y, w, h = cv2.boundingRect(contours[i])
if 2500 < w * h < 5e4 and h < 200 and w < 200:
roi_ptr.append(
(max(x - padding, 0), max(0, y - padding), min(w + 2 * padding, wi), min(hi, h + 2 * padding)))
cnt += 1
cv2.rectangle(img_tmp, (x - padding, y - padding), (x + w + padding, y + w + padding), (0, 0, 255), 2)
# for x, y, w, h in roi_ptr:
# print find_markers(img[y:y + w, x:x + w])
if len(roi_ptr) == 0:
return None, None, None
x, y, w, h = roi_ptr[0]
color, shape = find_markers(img[y:y + w, x:x + w])
return color, shape, len(roi_ptr)
def LabelDetector(self, AreaValue=2000, extendHeight=10, extendWight=10):
th = cv2.adaptiveThreshold(self.img, 255,
cv2.ADAPTIVE_THRESH_GAUSSIAN_C,
cv2.THRESH_BINARY_INV, 11, 2)
kernel = cv2.getStructuringElement(cv2.MORPH_RECT, (3, 3))
mol = cv2.morphologyEx(th, cv2.MORPH_OPEN, kernel, iterations=1)
mol = cv2.morphologyEx(mol, cv2.MORPH_CLOSE, kernel, iterations=7)
edge = imu.auto_canny(mol)
edge = cv2.morphologyEx(edge, cv2.MORPH_DILATE, kernel)
cnts = cv2.findContours(edge, cv2.RETR_EXTERNAL,
cv2.CHAIN_APPROX_SIMPLE)
cnts = imu.grab_contours(cnts)
if len(cnts) != 0:
(cnts, _) = contours.sort_contours(cnts)
for cnt in cnts:
area = cv2.contourArea(cnt)
if area > AreaValue:
box = cv2.minAreaRect(cnt)
box = cv2.boxPoints(box)
box = perspective.order_points(box)
box = np.array(box, dtype="int")
[tl, bl, br, tr] = self.cal_bouding_box(box)
maxWidth, maxHeight = self.getOjectSize(tl, bl, br, tr)
ratio = round(maxWidth / maxHeight, 4)
return box, ratio, edge, area
else:
pass
def auto_canny(self, sigma=0.33):
"""compute the median of the single channel pixel intensities"""
self.mat = imutils.auto_canny(
image=self.mat,
sigma=sigma,
)
return self
def get_contours(self, base_img):
"""得到图片四角"""
# 处理优化图片
h = cv2.Sobel(base_img, cv2.CV_32F, 0, 1, -1)
v = cv2.Sobel(base_img, cv2.CV_32F, 1, 0, -1)
img = cv2.add(h, v)
img = cv2.convertScaleAbs(img)
img = cv2.GaussianBlur(img, (3, 3), 0)
ret, img = cv2.threshold(img, 120, 255, cv2.THRESH_BINARY)
kernel = np.ones((3, 3), np.uint8)
# 膨胀
img = cv2.dilate(img, kernel, iterations=2)
img = auto_canny(img)
# 获取最大轮廓和轮廓周长
cnt, cnt_perimeter = self.get_max_area_cnt(img)
base_img_perimeter = (base_img.shape[0] + base_img.shape[1]) * 2
# 答题卡框与整个图片周长比的阈值
CNT_PERIMETER_THRESHOLD = 0.35
if not cnt_perimeter > CNT_PERIMETER_THRESHOLD * base_img_perimeter:
logger.error("[get_contours] 获取答题卡失败，请重新上传图片")
raise ImageException("获取答题卡失败，请重新上传图片")
# 10%，即0.1的精确度,忽略细节
epsilon = 0.1 * cv2.arcLength(cnt, True)
# 计算多边形的顶点，并看是否是四个顶点
poly_node_list = cv2.approxPolyDP(cnt, epsilon, True)
if not poly_node_list.shape[0] == 4:
logger.error("[get_contours] 不支持该答题卡，请重新上传图片")
raise ImageException("不支持该答题卡，请重新上传图片")
return poly_node_list
def canny(mat):
img = imutils.auto_canny(image=mat, sigma=0.3)
img = add_frame_labels(
frame=img,
labels=[f"canny cost: {canny.cost:6.3f}s"],
color=colors.get("WHITE"),
)
return img
def load_template(image, load_type):
try:
template = cv2.imread(f"{directory}/data/templates/{image}", load_type)
template = imutils.auto_canny(template)
except cv2.error:
print("Input one of the names in the 'data/templates' folder in order to detect")
return None
return template
def canny(self, visualization=False):
from imutils import auto_canny
# 보통은 영상처리전 원본이미지를 src(source), 목표로 하는 영상처리 결과이미지를 dst(destination)라고 네이밍해
src = self.get_image()
dst = auto_canny(src)
if visualization:
cv2.imshow("Canny", dst)
cv2.waitKey(1)
return dst
def extract_hog_histogram(image, bins=(8, 8, 8)):
gray = cv2.cvtColor(image, cv2.COLOR_BGR2GRAY)
edged = imutils.auto_canny(gray)
H = feature.hog(edged,
orientations=9,
pixels_per_cell=(10, 10),
cells_per_block=(2, 2),
normalise=True)
return H
def canny_edge_detection(frame: np.array, sigma=33) -> np.array:
"""
Automatic edge detection from the imutils library.
:param frame: A frame.
:param sigma: A magic number. Shouldn't be changed without testing.
:return: A black and white image of the edges in the frame.
"""
gray = cv2.cvtColor(frame, cv2.COLOR_BGR2GRAY)
return imutils.auto_canny(gray, sigma=sigma)
def _reset(self):
self.cost_matrix = None
self.contours = []
self.paths = []
self.count = 0 # number of closed nodes
self.max_level = 0 # maximum level of nodes, used in displaying path tree
self.seed = None
self.pq = None
self.pixel_map = None
self.edges = imutils.auto_canny(self.img)
def grab_screen():
try:
screenshot(f"{directory}/data/screenshots/screen_shot.JPG")
image = cv2.imread(f"{directory}/data/screenshots/screen_shot.JPG")
image = cv2.cvtColor(image, cv2.COLOR_BGR2GRAY)
image = imutils.auto_canny(image)
return image
except cv2.error:
print("Make sure you have 'data/screenshots' folder")
return None
return image
def get_err(frame):
# frame[-speed_patch.shape[0]:, -speed_patch.shape[1]:] = speed_patch
frame[-250:, :250] = map_patch
frame = frame[int(frame.shape[0] / 1.6):-130, 400:-400]
h, w, _ = frame.shape
frame = cv2.GaussianBlur(frame, initial_blur_kernel, 0)
frame = imutils.auto_canny(frame)
frame = er_cycle(frame, 5)
# cv2.imshow("w",frame)
def foo(r):
tmp = np.where(r > 0)
if tmp[0].shape[0] == 0:
return 0
else:
return int(np.mean(tmp))
centers = np.apply_along_axis(foo, 1, frame)
new_frame = np.copy(frame)
new_frame.fill(0)
for y in range(centers.shape[0]):
if centers[y] != 0:
cv2.circle(new_frame, (centers[y], y), 1, 255, -1)
kernel = np.ones((5, 5), np.uint8)
new_frame = cv2.dilate(new_frame, kernel, iterations=1)
new_frame = cv2.erode(new_frame, kernel, iterations=1)
frame = new_frame
points = list(zip(*np.where(frame > 0)[::-1]))
# print(len(points))
if len(points) < 2:
err = 0
frame = cv2.cvtColor(frame, cv2.COLOR_GRAY2RGB)
else:
[vx, vy, x, y] = cv2.fitLine(np.array(points), cv2.DIST_WELSCH, 0,
0.01, 0.01)
m = 100
x0, y0 = int(x[0] - m * vx[0]), int(y[0] - m * vy[0])
x1, y1 = int(x[0] + m * vx[0]), int(y[0] + m * vy[0])
err_y = h - 1
if y1 == y0:
err_x = x1
else:
err_x = x1 + (err_y - y1) * (x0 - x1) / (y0 - y1)
err = err_x - w // 2
frame = cv2.cvtColor(frame, cv2.COLOR_GRAY2RGB)
frame = cv2.line(frame, (x0, y0), (x1, y1), (255, 0, 0), 5)
cv2.putText(frame, 'key: %s' % key, (10, h - 50), cv2.FONT_HERSHEY_SIMPLEX,
1, (255, 255, 255), 2)
cv2.putText(frame, 'err: %d' % err, (10, h - 25), cv2.FONT_HERSHEY_SIMPLEX,
1, (255, 255, 255), 2)
out.write(frame)
return err
def Contours(gray_image, print_true=True, test=False):
#temp = cv2.Canny(gray_image, 50, 100)
temp = imutils.auto_canny(gray_image)
if test: cv2.imshow("Canny A", temp)
else:
temp = cv2.dilate(temp, None, iterations=1)
temp = cv2.erode(temp, None, iterations=1)
if test: cv2.imshow("Canny B", temp)
if (print_true): cv2.imshow("EdgeDETECT", temp)
cnts = cv2.findContours(temp.copy(), cv2.RETR_EXTERNAL,
cv2.CHAIN_APPROX_SIMPLE)
cnts = imutils.grab_contours(cnts)
(cnts, _) = contours.sort_contours(cnts)
return cnts
def process_image(raw, mark_image=False):
cannyd = imutils.auto_canny(raw)
height, width = cannyd.shape
marker_width = coords['width']
marker_halfwidth = int(marker_width * .5)
font = cv2.FONT_HERSHEY_SIMPLEX
fontScale = .5
color = (255, 0, 0)
thickness = 2
spaces_left = len(coords['coords'])
for coord in coords['coords']:
x = int(coord['x']) - marker_width
y = int(coord['y']) - marker_halfwidth
x_max = x + marker_width
y_max = y + marker_width
if x_max > width:
x_max = width
if y_max > height:
y_max = height
if y < 0:
y = 0
if x < 0:
x = 0
box = cannyd[y:y_max, x:x_max]
n_white_pix = np.sum(box == 255)
percent_white = round(n_white_pix / marker_halfwidth**2, 2)
if percent_white > .1:
spaces_left -= 1
if mark_image:
start_point = (x, y)
end_point = (x_max, y_max)
cannyd = cv2.rectangle(cannyd, start_point, end_point, (255, 0, 0),
2)
cannyd = cv2.putText(cannyd, str(percent_white),
(x, y - marker_halfwidth), font, fontScale,
color, thickness, cv2.LINE_AA)
return cannyd, spaces_left
def auto_canny(img_path):
img = cv2.imread(img_path, 0)
img_canny = imutils.auto_canny(img)
img_skeleton = imutils.skeletonize(img, size=(3, 3))
plt.subplot(131)
plt.imshow(imutils.opencv2matplotlib(img_canny))
plt.axis('off')
plt.subplot(132)
plt.imshow(imutils.opencv2matplotlib(img))
plt.axis('off')
plt.subplot(133)
plt.imshow(imutils.opencv2matplotlib(img_skeleton))
plt.axis('off')
plt.show()
def apply_filter(img, select="Gray"):
filters = {}
output = img.copy()
filters["Original"] = output
filters["Blur"] = cv2.GaussianBlur(output, (3, 3), 0)
filters["Gray"] = cv2.cvtColor(filters["Blur"], cv2.COLOR_BGR2GRAY)
filters["Thresh"] = cv2.threshold(
filters["Gray"], 113, 255, cv2.THRESH_BINARY)[1] # 60
filters["Thresh-Adaptive"] = cv2.adaptiveThreshold(
filters["Gray"], 255, cv2.ADAPTIVE_THRESH_GAUSSIAN_C, cv2.THRESH_BINARY, 11, 2)
filters["Canny"] = cv2.Canny(filters["Blur"], 100, 200)
filters["Canny-Auto"] = imutils.auto_canny(filters["Blur"])
return filters[select]
def lane_detection_pipeline(img,
gaussian_kernel_size=3,
rho=1,
theta=np.pi / 180.0,
threshold=40,
min_line_len=40,
max_line_gap=80):
result = grayscale(img)
result = gaussian_blur(result, gaussian_kernel_size)
result = auto_canny(result)
result = region_of_interest(result, create_region_of_interest(img.shape))
result = cv2.equalizeHist(result)
lines_image = hough_lines(result, rho, theta, threshold, min_line_len,
max_line_gap)
return weighted_img(lines_image, img)
def extract_image_features(self, data):
# Please do not modify the header above
# extract feature vector from image data
feature_data = []
for img in data:
#gray = color.rgb2gray(img)
edged = imutils.auto_canny(img)
(H, hogImage) = feature.hog(edged, orientations=9, pixels_per_cell=(90, 90), cells_per_block=(2, 2), transform_sqrt=True, visualise=True)
#viewer = ImageViewer(hogImage)
#viewer.show()
feature_data.append(H)
return(feature_data)
def _create_edge_map(self, sigma=0.33, dilate_kernel_size=7):
"""Convert image into dilated edge map
:param sigma: value to be passed to imutils.auto_canny
:param dilate_kernel_size: kernel radius to be used during dilation step
:return: numpy array/opencv gray image containing dilated edge map
"""
edge_map = imutils.auto_canny(self._background, sigma=sigma)
if dilate_kernel_size > 2:
if dilate_kernel_size % 2 == 0:
dilate_kernel_size -= 1
kernel = cv2.getStructuringElement(
cv2.MORPH_ELLIPSE, (dilate_kernel_size, dilate_kernel_size))
edge_map = cv2.dilate(edge_map, kernel)
return edge_map
def displayFilters(self, select="Gray"):
filters = {}
output = self.frame
filters["Original"] = output
filters["Blur"] = cv2.GaussianBlur(output, (7, 7), 0)
filters["Gray"] = cv2.cvtColor(filters["Blur"], cv2.COLOR_BGR2GRAY)
filters["Thresh"] = cv2.threshold(filters["Gray"], 145, 255,
cv2.THRESH_BINARY)[1] # 60
filters["Thresh-Adaptive"] = cv2.adaptiveThreshold(
filters["Gray"], 255, cv2.ADAPTIVE_THRESH_GAUSSIAN_C,
cv2.THRESH_BINARY, 5, 1)
#filters["Thresh-Adaptive"] = cv2.dilate(filters["Thresh-Adaptive"], None, iterations=0)
filters["Canny"] = cv2.Canny(filters["Blur"], 100, 200)
filters["Canny-Auto"] = imutils.auto_canny(filters["Blur"])
cv2.imshow("Camera view:", filters[select])
def canny_edge(image):
gray = cv2.cvtColor(image, cv2.COLOR_BGR2GRAY)
blurred = cv2.GaussianBlur(gray, (5, 5), 0)
cv2.imshow("Gray", gray)
cv2.imshow("Blurred", blurred)
# compute a "wide", "mid-range", and "tight" threshold for the edges
wide = cv2.Canny(blurred, 10, 200)
mid = cv2.Canny(blurred, 30, 150)
tight = cv2.Canny(blurred, 245, 250)
auto_range = imutils.auto_canny(blurred)
cv2.imshow("Wide Edge Map", wide)
cv2.imshow("Mid Edge Map", mid)
cv2.imshow("Tight Edge Map", tight)
cv2.imshow("Auto canny", auto_range)
def canny_edge(image):
gray = cv2.cvtColor(image, cv2.COLOR_BGR2GRAY)
blurred = cv2.GaussianBlur(gray, (5, 5), 0)
cv2.imshow("Gray", gray)
cv2.imshow("Blurred", blurred)
# compute a "wide", "mid-range", and "tight" threshold for the edges
wide = cv2.Canny(blurred, 10, 200)
mid = cv2.Canny(blurred, 30, 150)
tight = cv2.Canny(blurred, 245, 250)
auto_range = imutils.auto_canny(blurred)
cv2.imshow("Wide Edge Map", wide)
cv2.imshow("Mid Edge Map", mid)
cv2.imshow("Tight Edge Map", tight)
cv2.imshow("Auto canny", auto_range)
def trainData():
print('training Started...')
data = []
labels = []
count = 0
dataFile = open('data.txt', 'w')
labelFile = open('labels.txt', 'w')
for car in vehicleList[:40000]:
# extract the make of the car
make = car.get('carType')
imagePath = car.get('imagePath')
# load the image, convert it to grayscale, and detect edges
image = cv2.imread(imagePath)
gray = cv2.cvtColor(image, cv2.COLOR_BGR2GRAY)
edged = imutils.auto_canny(gray)
# find contours in the edge map, keeping only the largest one which
# is presmumed to be the car logo
cnts = cv2.findContours(edged.copy(), cv2.RETR_EXTERNAL,
cv2.CHAIN_APPROX_SIMPLE)
cnts = cnts[0] if imutils.is_cv2(cnts) else cnts[1]
c = max(cnts, key=cv2.contourArea)
# extract the logo of the car and resize it to a canonical width
# and height
(x, y, w, h) = cv2.boundingRect(c)
logo = gray[y:y + h, x:x + w]
logo = cv2.resize(logo, (200, 100))
# extract Histogram of Oriented Gradients from the logo
H = feature.hog(logo, orientations=9, pixels_per_cell=(10, 10),
cells_per_block=(2, 2), transform_sqrt=True, block_norm="L1")
# update the data and labels
data.append(H)
dataFile.write(str(H))
dataFile.write("\n")
labels.append(make)
labelFile.write(str(make))
labelFile.write("\n")
count += 1
return data, labels
def Predict(image, model):
gray = cv2.cvtColor(image, cv2.COLOR_BGR2GRAY)
edge = imutils.auto_canny(gray)
resized = cv2.resize(gray, image_resize)
H, hogImage = feature.hog(resized,
orientations=9,
pixels_per_cell=ppc,
cells_per_block=cpb,
transform_sqrt=True,
block_norm="L1",
visualize=True)
pred = model.predict(H.reshape(1, -1))[0]
hogImage = exposure.rescale_intensity(hogImage, out_range=(0, 255))
hogImage = hogImage.astype("uint8")
#cv2.imshow("Hog Image No. {}".format(i+1), hogImage)
cv2.putText(image, pred.title(), (10, 35), cv2.FONT_HERSHEY_SIMPLEX, 1.0,
(110, 34, 59), 3)
cv2.imshow("Test Image No. {}".format(i + 1), image)
def edges(self, frame=None, **params):
rs = None
image = self.frame(frame, **params)
if not image is None:
gray = cv2.cvtColor(image, cv2.COLOR_BGR2GRAY)
blurred = cv2.GaussianBlur(gray, (3, 3), 0)
# apply Canny edge detection using a wide threshold, tight
# threshold, and automatically determined threshold
# edged = cv2.Canny(blurred, 10, 200)
# edged = cv2.Canny(blurred, 225, 250)
edged = imutils.auto_canny(blurred)
rs = {
'frame': image,
'iframe': gray,
'oframe': edged,
}
return rs
def Edge_Canny_auto(self):
# read image
im = cv2.imread(self.Image)
gray = cv2.cvtColor(im, cv2.COLOR_BGR2GRAY)
blurred = cv2.GaussianBlur(gray, (3, 3), 0)
# apply Canny edge detection using a wide threshold, tight
# threshold, and automatically determined threshold
wide = cv2.Canny(blurred, 10, 200)
tight = cv2.Canny(blurred, 225, 250)
auto = imutils.auto_canny(blurred)
# show the images
cv2.imshow("Original", im)
cv2.imshow("Wide", wide)
cv2.imshow("Tight", tight)
cv2.imshow("Auto", auto)
cv2.waitKey(0)
def Edge_Canny_auto(self):
# read image
im = cv2.imread(self.Image)
gray = cv2.cvtColor(im, cv2.COLOR_BGR2GRAY)
blurred = cv2.GaussianBlur(gray, (3, 3), 0)
# apply Canny edge detection using a wide threshold, tight
# threshold, and automatically determined threshold
wide = cv2.Canny(blurred, 10, 200)
tight = cv2.Canny(blurred, 225, 250)
auto = imutils.auto_canny(blurred)
# show the images
cv2.imshow("Original", im)
cv2.imshow("Wide", wide)
cv2.imshow("Tight", tight)
cv2.imshow("Auto", auto)
cv2.waitKey(0)
def approx_realworld():
image = cv2.imread("../img/receipt.png")
gray = cv2.cvtColor(image, cv2.COLOR_BGR2GRAY)
edged = imutils.auto_canny(gray, 3.33)
cv2.imshow("Original", image)
cv2.imshow("Edge", edged)
(_, cnts, _) = cv2.findContours(edged.copy(), cv2.RETR_EXTERNAL, cv2.CHAIN_APPROX_SIMPLE)
cnts = sorted(cnts, key=cv2.contourArea, reverse=True)[:7]
for c in cnts:
peri = cv2.arcLength(c, True)
approx = cv2.approxPolyDP(c, 0.01 * peri, True)
print("original: {}, approx: {}".format(len(c), len(approx)))
if len(approx) == 4:
cv2.drawContours(image, [approx], -1, (0, 255, 0), 3)
cv2.imshow("Image", image)
cv2.imshow("Skeleton", skeleton)
cv2.waitKey(0)
cv2.destroyAllWindows()
# 5. MATPLOTLIB
# INCORRECT: show the image without converting color spaces
plt.figure("Incorrect")
plt.imshow(cactus)
# CORRECT: convert color spaces before using plt.imshow
plt.figure("Correct")
plt.imshow(imutils.opencv2matplotlib(cactus))
plt.show()
# 6. URL TO IMAGE
# load an image from a URL, convert it to OpenCV, format, and
# display it
url = "http://pyimagesearch.com/static/pyimagesearch_logo_github.png"
logo = imutils.url_to_image(url)
cv2.imshow("URL to Image", logo)
cv2.waitKey(0)
cv2.destroyAllWindows()
# 7. AUTO CANNY
# convert the logo to grayscale and automatically detect edges
gray = cv2.cvtColor(logo, cv2.COLOR_BGR2GRAY)
edgeMap = imutils.auto_canny(gray)
cv2.imshow("Original", logo)
cv2.imshow("Automatic Edge Map", edgeMap)
cv2.waitKey(0)
image = ndimage.imread( "..\\Exemple-annonces\\ordi7.jpg") plt.imshow(image) plt.show() hsv = cv2.cvtColor(image, cv2.COLOR_BGR2HSV) lower = np.array([115, 135, 135]) #RGB upper = np.array([125, 255, 255]) #RGB red = cv2.inRange(hsv, lower, upper) gray = cv2.cvtColor(image, cv2.COLOR_BGR2GRAY) plt.imshow(gray,cmap='gray') plt.show() edged = imutils.auto_canny(red) plt.imshow(edged,cmap='gray') plt.show() im2,contours, hierarchy = cv2.findContours(edged.copy(), cv2.RETR_EXTERNAL,cv2.CHAIN_APPROX_SIMPLE) contours.sort(key=cv2.contourArea,reverse=True ) area = [cv2.contourArea(contours[i]) for i in range(len(contours)) ] (x, y, w, h) = cv2.boundingRect(contours[1]) logo = gray[y:y + h, x:x + w] plt.imshow(logo) plt.show() H = feature.hog(logo, orientations=9, pixels_per_cell=(10, 10),
# USAGE
# BE SURE TO INSTALL 'imutils' PRIOR TO EXECUTING THIS COMMAND
# python sorting_contours.py
# import the necessary packages
from imutils import contours
import imutils
import cv2
# load the shapes image clone it, convert it to grayscale, and
# detect edges in the image
image = cv2.imread("../demo_images/shapes.png")
orig = image.copy()
gray = cv2.cvtColor(image, cv2.COLOR_BGR2GRAY)
edged = imutils.auto_canny(gray)
# find contours in the edge map
(cnts, _) = cv2.findContours(edged.copy(), cv2.RETR_EXTERNAL,
cv2.CHAIN_APPROX_SIMPLE)
# loop over the (unsorted) contours and label them
for (i, c) in enumerate(cnts):
orig = contours.label_contour(orig, c, i, color=(240, 0, 159))
# show the original image
cv2.imshow("Original", orig)
# loop over the sorting methods
for method in ("left-to-right", "right-to-left", "top-to-bottom", "bottom-to-top"):
# sort the contours
