def detect_by_gf(origin,
isdebug=False):
if origin is None:
return None
# Default Size
h,w,c = origin.shape
size = 200.0
# Resize
img = cv2.resize(origin,(int(w*size/h),int(size)))
# Extract Good Features
corners = refinedGoodFeatures(origin,img,
model='LP')
mask = checkFeatures(img,corners,True)
# Find Candidate
bboxes = findBBox(img,mask,
model='LP',
debug=True)
# Resize
if bboxes is not None:
bboxes = resizeBBoxes(bboxes,h/size)
# Check Result
if isdebug and bboxes is not None:
drawBBox(origin,bboxes,debug=True)
# Crop Rois
rois = BBoxes2ROIs(origin,bboxes)
if isdebug and rois is not None:
for i in range(len(rois)):
showResult("cropped",rois[i])
return bboxes,rois
def colormask(image,
isday=True,
isdebug=False):
# blue mask + yellow mask
hsv = cv2.cvtColor(image,cv2.COLOR_BGR2HSV)
h,s,v = cv2.split(hsv)
sat = (s.copy()).astype('float')
val = (v.copy()).astype('float')
sat /= 255
val /= 255
#
blue = h.copy()
yellow = h.copy()
if isday:
blue = np.where((h<100)|(h>135)|(v<40)|(s<40),0,255)
yellow = np.where((h<20)|(h>40)|(v<40)|(s<40),0,255)
else:
blue = np.where((h<105)|(h>135),0,255)
yellow = np.where((h<20)|(h>40),0,255)
#
hue = np.zeros(image.shape[:2])
hue[hue>=0] = 0.2
hue[blue>0] = 0.9
hue[yellow>0] = 0.7
if isday:
mask=(hue*sat*val)**2
else:
mask=(hue*sat)**3
#
if isdebug:
showResult("colormask:mask",mask*255)
return mask
def find_possible_chars(thr, isdebug=False):
#
tmp = thr.copy()
img_contour, contours, npaHierarchy = cv2.findContours(
tmp, cv2.RETR_LIST, cv2.CHAIN_APPROX_SIMPLE) # find all contours
height, width = thr.shape
img_contour = np.zeros((height, width, 3), np.uint8)
#
filtered = [] # this will be the return value
count = 0
for i in range(0, len(contours)): # for each contour
if isdebug:
cv2.drawContours(img_contour, contours, i, SCALAR_WHITE)
contour = Contour(contours[i])
if contour.checkIfPossibleChar(
): # if contour is a possible char, note this does not compare to other chars (yet) . . .
count += 1 # increment count of possible chars
filtered.append(contour) # and add to list of possible chars
if isdebug:
print "\nstep 2 - contours = " + str(
len(contours)) # 2362 with MCLRNF1 image
print "step 2 - chars = " + str(count) # 131 with MCLRNF1 image
showResult("contours", img_contour)
return filtered
def tophatblackhat(gray):
gray = cv2.GaussianBlur(gray,(3,3),0)
rectKernel = cv2.getStructuringElement(cv2.MORPH_RECT, (9, 3))
tophat = cv2.morphologyEx(gray, cv2.MORPH_TOPHAT, rectKernel)
blackhat = cv2.morphologyEx(gray, cv2.MORPH_BLACKHAT, rectKernel)
showResult("tophat",tophat)
showResult("blackhat",blackhat)
def removeshadow(img):
#
hsv = cv2.cvtColor(img, cv2.COLOR_BGR2HSV)
gray = cv2.cvtColor(img, cv2.COLOR_BGR2GRAY)
gray = 255 - gray
hsv[:, :, 2] = gray
bgr = cv2.cvtColor(hsv, cv2.COLOR_HSV2BGR)
#
rgb_planes = cv2.split(bgr)
result_planes = []
result_norm_planes = []
for plane in rgb_planes:
dilated_img = cv2.dilate(plane, np.ones((7, 7), np.uint8))
bg_img = cv2.medianBlur(dilated_img, 21)
diff_img = 255 - cv2.absdiff(plane, bg_img)
norm_img = diff_img.copy()
norm_img = cv2.normalize(diff_img,
norm_img,
alpha=0,
beta=255,
norm_type=cv2.NORM_MINMAX,
dtype=cv2.CV_8UC1)
result_planes.append(diff_img)
result_norm_planes.append(norm_img)
result = cv2.merge(result_planes)
result_norm = cv2.merge(result_norm_planes)
showResult("result", result)
showResult("result_norm", result_norm)
return result, result_norm
def SimpleRecognizePlate(image):
t0 = time.time()
bboxes, images = detect_by_probability(
image) #detect.detectPlateRough(image)
if images is None:
return
for image in images:
#image = cv2.cvtColor(image, cv2.COLOR_RGB2GRAY)
image = cv2.resize(image, (136, 36))
image = kojy_gray(image)
#image = maximizeContrast(255 - image)
image_gray = fm.findContoursAndDrawBoundingBox(image)
showResult("plate-colr", image)
showResult("plate-gray,", image_gray)
blocks, res, confidence = segmentation.slidingWindowsEval(image_gray)
#for i in range(len(blocks)):
# showResult("plate-gray,",blocks[i])
if confidence > 4.5:
print "车牌:", res, "置信度:", confidence
else:
print "不确定的车牌:", res, "置信度:", confidence
print time.time() - t0, "s"
def setcontours(self,isdebug=False):
#self.img_contour, self.contours, npaHierarchy = cv2.findContours(self.compose, cv2.RETR_LIST, cv2.CHAIN_APPROX_SIMPLE) # find all contours
self.contours, npaHierarchy = cv2.findContours(self.compose, cv2.RETR_LIST, cv2.CHAIN_APPROX_SIMPLE) # find all contours
self.img_contour = self.compose
if isdebug:
cv2.drawContours(self.img_contour, self.contours, -1, SCALAR_WHITE, -1)
print "\nstep 2 - contours = " + str(len(self.contours))
showResult("contours",self.img_contour)
def mask2plates(img, finalmask, isdebug=False):
#showResult("masktest",finalmask)
#ret,binary = cv2.threshold(finalmask,0,255,cv2.THRESH_BINARY+cv2.THRESH_OTSU)\
binary = compositeThreshold(finalmask, mode='otsu')
closing = close(binary)
if isdebug:
showResult("mask2plates:masktest", closing)
# Find Candidate
return findBBox(img, closing, isdebug=isdebug)
def compositeThreshold(gray, mode='com'):
if mode == 'otsu':
otsu = threshold_otsu(gray)
otsu_bin = gray > otsu
otsu_bin = otsu_bin.astype(np.uint8) * 255
return otsu_bin
elif mode == 'yen':
yen = threshold_yen(gray)
yen_bin = gray > yen
yen_bin = yen_bin.astype(np.uint8) * 255
return yen_bin
elif mode == 'li':
li = threshold_li(gray)
li_bin = gray > li
li_bin = li_bin.astype(np.uint8) * 255
return li_bin
elif mode == 'niblack':
niblack = threshold_niblack(gray, window_size=13, k=0.8)
niblack_bin = gray > niblack
niblack_bin = niblack_bin.astype(np.uint8) * 255
return niblack_bin
elif mode == 'sauvola':
sauvola = threshold_sauvola(gray, window_size=13)
sauvola_bin = gray > sauvola
sauvola_bin = sauvola_bin.astype(np.uint8) * 255
return sauvola_bin
elif mode == 'com':
li = threshold_li(gray)
li_bin = gray > li
li_bin = li_bin.astype(np.uint8) * 255
otsu = threshold_otsu(gray)
otsu_bin = gray > otsu
otsu_bin = otsu_bin.astype(np.uint8) * 255
yen = threshold_yen(gray)
yen_bin = gray > yen
yen_bin = yen_bin.astype(np.uint8) * 255
return cv2.min(cv2.min(otsu_bin, li_bin), yen_bin)
elif mode == "niblack-multi":
thr = np.zeros((gray.shape), dtype=np.uint8)
thr[thr >= 0] = 255
for k in np.linspace(-0.8, 0.2, 5): #(-1.8,0.2,5)
thresh_niblack = threshold_niblack(gray, window_size=25, k=k)
binary_niblack = gray > thresh_niblack
binary_niblack = binary_niblack.astype(np.uint8) * 255
showResult("binary_niblack", binary_niblack)
thr = cv2.min(thr, binary_niblack)
return thr
else:
sauvola = threshold_sauvola(gray, window_size=25, k=0.25)
sauvola_bin = gray > sauvola
sauvola_bin = sauvola_bin.astype(np.uint8) * 255
niblack = threshold_niblack(gray, window_size=25, k=0.25)
niblack_bin = gray > niblack
niblack_bin = niblack_bin.astype(np.uint8) * 255
return cv2.max(sauvola, niblack)
def drawChars(img, title, contours, colr=SCALAR_WHITE, isdebug=False):
img_contours = np.zeros((img.shape), np.uint8)
contours_ = []
for contour in contours:
contours_.append(contour.contour)
cv2.drawContours(img_contours, contours_, -1, colr)
if isdebug:
showResult(title, img_contours)
return img_contours
def colormask(image, isday=True, tgtcolr='Default', isdebug=False):
#
#image = NormalizeT(image)
#showResult("normalized",image)
# blue mask + yellow mask
hsv = cv2.cvtColor(image, cv2.COLOR_BGR2HSV)
h, s, v = cv2.split(hsv)
sat = (s.copy()).astype('float')
val = (v.copy()).astype('float')
sat /= 255
val /= 255
#
hue = np.zeros(image.shape[:2])
hue[hue >= 0] = 0.2
#
blue1 = h.copy()
blue2 = h.copy()
yellow = h.copy()
#
if isday:
blue1 = np.where((h < 105) | (h > 135) | (v < 40) | (s < 40), 0, 255)
blue2 = np.where((h < 100) | (h > 140) | (v < 30) | (s < 30), 0, 255)
yellow = np.where((h < 20) | (h > 40) | (v < 40) | (s < 40), 0, 255)
else:
blue1 = np.where((h < 105) | (h > 135), 0, 255)
blue2 = np.where((h < 100) | (h > 140), 0, 255)
yellow = np.where((h < 20) | (h > 40), 0, 255)
#
for case in switch(tgtcolr):
if case('Blue'):
hue[blue2 > 0] = 0.5
hue[blue1 > 0] = 1.0
hue[yellow > 0] = 0.1
break
if case('Yellow'):
hue[yellow > 0] = 1.0
hue[blue2 > 0] = 0.1
hue[blue1 > 0] = 0.15
break
if case('Default'):
hue[blue2 > 0] = 0.4
hue[blue1 > 0] = 0.9
hue[yellow > 0] = 0.7
break
#
if isday:
mask = (hue * sat * val)**2 #(hue*sat*val)**2
else:
mask = (hue * sat)**2
#
if isdebug:
showResult("colormask:mask", mask)
return mask
def detect_by_seg_gf(origin,
isdebug=False):
if origin is None:
return None
# Default Size
h,w,c = origin.shape
size = 200.0
#origin = opencv2skimage(origin)
# Resize
img = cv2.resize(origin,(int(w*size/h),int(size)))
# Blur
blur = cv2.GaussianBlur(img,(5,5),3)
# Equalization Hist
#origin = equalizehist(origin)
# Extract Good Features
corners = refinedGoodFeatures(origin,img,
model='LP')
corners_,handwrite = refinedCorners(img,corners,False)
checkFeatures(img,corners,True)
# Opencv2Skimage
skimg = cv2.cvtColor(blur, cv2.COLOR_BGR2RGB)
# Segmentation
out,labels = seg(skimg,Debug=False)
# Eval Label
labels = refineLabels(out,labels,corners_,howmany=20)
# Show Result
out = drawLabels(skimg,labels,Debug=isdebug)
if out is None:
return None
changebgcolr(out,labels)
#showResult("labelout",skimage2opencv(out))
# Find Candidate
bboxes = findBBox(img,out,
model='LP',
debug=isdebug)
# Resize
if bboxes is not None:
bboxes = resizeBBoxes(bboxes,h/size)
# Check Candidate
if isdebug and bboxes is not None:
drawBBox(origin,bboxes,debug=isdebug)
# Crop Rois
rois = BBoxes2ROIs(origin,bboxes)
if isdebug and rois is not None:
for i in range(len(rois)):
showResult("cropped",rois[i])
'''
if isdebug:
#labels2boundaries(labels)
contours = labels2contours(labels)
drawBBox(img,contours,debug=True)
'''
return bboxes,rois
def mkfinalmask(image, gf_image=None, isday=True, isdebug=False):
colrmask = colormask(image, isday)
regmask = regionmask(image)
mask = colrmask * regmask
if gf_image is not None:
gfmask = featuremask(gf_image)
mask *= gfmask
mask = (mask * 255).astype('uint8')
if isdebug:
showResult("mkfinalmask:mask", mask * 255)
return mask
def finalize(self, isdebug=False):
strings_ = self.makeup(isdebug=isdebug)
# need ocr or not
strings_ = VIN.ocrchecking(self.bgr, strings_, False)
#
if strings_ is not None:
marked = String.mark(self.bgr, strings_)
if isdebug:
showResult("marked", marked)
#drawChars((self.height,self.width),"final",string.getitems(),isdebug=isdebug)
#print string.result
return True, strings_[0].confidence, marked
else:
return False, 0.0, None
def mkfinalmasks(image, gf_image=None, isday=True, isdebug=False):
masks = []
regmask = regionmask(image)
for i in range(2):
colrmask = colormask(image, isday, tgtcolr=colrs[i], isdebug=isdebug)
mask = colrmask * regmask
if gf_image is not None:
gfmask = featuremask(gf_image)
mask *= gfmask
mask = (mask * 255).astype('uint8')
if isdebug:
showResult("mkfinalmask:mask", mask)
masks.append(mask)
return masks
def makesegments(image,string):
string.ROI = String.extractROI(image,string,False)
bgr = string.ROI.cropped
chars = string.chars
FirstIndex = 0
LastIndex = string.charcount - 1
h,w = bgr.shape[:2]
delta = w - (chars[LastIndex].brX + chars[LastIndex].brW - chars[FirstIndex].brX)
first = chars[0].brX
for char in chars:
x = char.brX - first + delta
segment = bgr[1:string.charheight,x:x+string.charwidth,:]
string.segments.append(segment)
showResult("segment",segment)
def recognizeLP(gray):
t0 = time.time()
if gray is None:
return
gray = cv2.resize(gray, (136, 36))
blocks, res, confidence = segmentation.slidingWindowsEval(gray)
for i in range(len(blocks)):
showResult("plate-gray,", blocks[i])
if confidence > 4.5:
print "车牌:", res, "置信度:", confidence
else:
print "不确定的车牌:", res, "置信度:", confidence
print time.time() - t0, "s"
return blocks, res, confidence
def detect_by_cascade(origin, resize_h=720, en_scale=1.06, isdebug=False):
if origin is None:
return None
# Default Size
height, width, channels = origin.shape
# Resize
resized = cv2.resize(origin, (int(width * resize_h / height), resize_h))
# Colr2Gray
gray = cv2.cvtColor(resized, cv2.COLOR_RGB2GRAY)
# Detect by Cascade
watches = watch_cascade.detectMultiScale(gray,
en_scale,
1,
minSize=(36, 9))
cropped_images = []
bboxes = []
for (x, y, w, h) in watches:
xmin = x - w * 0.1
ymin = y - h * 0.6
xmin = xmin if xmin >= 0 else 0
xmin = ymin if ymin >= 0 else 0
xmax = xmin + 1.2 * w
ymax = ymin + 1.1 * h
bboxes.append(
np.array([[xmin, ymin], [xmax, ymin], [xmax, ymax], [xmin, ymax]]))
cropped = cropped_from_image(
gray, (int(xmin), int(ymin), int(1.2 * w), int(1.1 * h)))
cropped_images.append(cropped)
# Resize
if bboxes is not None:
bboxes = resizeBBoxes(bboxes, height / float(resize_h))
# Check Result
if isdebug and bboxes is not None:
drawBBox(origin, bboxes, debug=True)
# Crop Rois
rois = BBoxes2ROIs(origin, bboxes)
if isdebug and rois is not None:
for i in range(len(rois)):
showResult("cropped", rois[i])
return bboxes, rois #cropped_images
def detect_by_probability(origin,
isdebug=False):
if origin is None:
return None
# Default Size
h,w,c = origin.shape
size = 200.0
# Resize
img = cv2.resize(origin,(int(w*size/h),int(size)))
#showResult("img",img)
#
if dayornight(img):
# Extract Good Features
corners = refinedGoodFeatures(origin,img)
mask = checkFeatures(img,corners,False)
closing=close(mask)
refined_gfmask = refine_gfimage(img,closing)
#showResult("refined_gfmask",refined_gfmask)
finalmask = mkfinalmask(img,refined_gfmask,isday=True)
else:
finalmask = mkfinalmask(img,None,isday=False)
#
#showResult("masktest",finalmask)
#ret,binary = cv2.threshold(finalmask,0,255,cv2.THRESH_BINARY+cv2.THRESH_OTSU)
binary = compositeThreshold(finalmask,mode='otsu')
kernel = cv2.getStructuringElement(cv2.MORPH_RECT,(5,5))
closing=cv2.morphologyEx(binary, cv2.MORPH_CLOSE, kernel)
if isdebug:
showResult("masktest",closing)
# Find Candidate
bboxes = findBBox(img,closing,isdebug=True)
# Resize
if bboxes is not None:
bboxes = resizeBBoxes(bboxes,h/size)
# Check Result
if isdebug and bboxes is not None:
drawBBox(origin,bboxes,debug=True)
# Crop Rois
rois = BBoxes2ROIs(origin,bboxes)
if isdebug and rois is not None:
for i in range(len(rois)):
showResult("cropped",rois[i])
return bboxes,rois
def drawChars(shape, title, chars, colr=SCALAR_WHITE, isdebug=False):
h, w = shape
ratio = 1
vis = np.zeros((h, w * ratio, 3), np.uint8)
img_contours = np.zeros((shape), np.uint8)
for i, char in enumerate(chars):
cv2.drawContours(img_contours, [char.contour], -1, colr, -1)
if isdebug:
[x, y] = char.brX, char.brY
cv2.putText(vis, str(i), (x * ratio, y),
cv2.FONT_HERSHEY_COMPLEX_SMALL, 1, (255, 128, 128), 1,
cv2.LINE_AA)
contour = char.contour.copy()
contour[:, :, 0] *= ratio
cv2.drawContours(vis, [contour], -1, colr, -1)
if isdebug:
showResult(title, vis)
return img_contours
def refinedCorners(img,
corners,
isdebug=False):
#http://scikit-image.org/docs/dev/auto_examples/edges/plot_skeleton.html
mark = img.copy()
mark[mark>=0] = 0
#img = cv2.drawKeypoints(img,corners)
for i in corners:
x,y = i.ravel()
cv2.circle(mark,(int(x),int(y)),1,(255,255,255),5)
dilate = cv2.dilate(mark[:,:,0],(3,3),iterations=1)
erode = cv2.dilate(dilate,(3,3),iterations=1)
closing = morphological(erode,cv2.MORPH_CLOSE)
erode = cv2.dilate(closing,(3,3),iterations=1)
#erode = cv2.erode(closing,(3,3),iterations=1)
#ret,thr = cv2.threshold(erode,0,255,cv2.THRESH_BINARY)
binary = erode#mark[:,:,0]
binary[binary>0]=1
skeleton = img_as_ubyte(morphology.skeletonize_3d(binary))#skeletonize(binary))
skeleton[skeleton > 0] = 255
dilate = cv2.dilate(skeleton,(3,3))
corners_ = np.transpose(np.nonzero(skeleton))
cv_corners = []
for corner in corners_:
cv_corners.append([np.array([corner[1],corner[0]])])
if isdebug:
for pt in cv_corners:
cv2.circle(img,(pt[0][0],pt[0][1]),1,(255,255,255),1)
showResult("mark",mark)
showResult("skeleton",skeleton)
showResult("img",img)
return cv_corners,mark[:,:,0]
def first_filtering(self, isdebug=False):
img_contour, contours, npaHierarchy = cv2.findContours(
self.thr, cv2.RETR_LIST,
cv2.CHAIN_APPROX_SIMPLE) # find all contours
height, width = self.thr.shape
img_contour = np.zeros((height, width, 3), np.uint8)
cheights = []
cwidths = []
for contour in contours:
if isdebug:
cv2.drawContours(img_contour, [contour], -1, SCALAR_WHITE)
[x, y, w, h] = cv2.boundingRect(contour)
braRatio = float(w) / float(h) #aspect ratio
if (w*h > MIN_PIXEL_AREA and\
w > MIN_PIXEL_WIDTH and\
h > self.height*0.4 and\
0.04 < braRatio and\
braRatio < 1.25):
self.contours.append(contour)
cwidths.append(w)
cheights.append(h)
if len(cwidths) < 2:
return None
self.charheight = statistics.median_high(cheights)
self.charwidth = statistics.median_high(cwidths)
if isdebug:
print("\nfiltering step 1 - contours = " + str(len(contours)))
showResult("contours", img_contour)
return self.contours
def process(self, roi=None, mode="Blue", isdebug=False):
self.initialize()
if roi is not None:
self.originheight, self.originwidth = roi.shape[:2]
self.bgr = cv2.resize(roi, (self.width, self.height),
interpolation=cv2.INTER_CUBIC)
self.preprocess(mode=mode)
#
if isdebug:
self.showAll()
# filtering contours
if self.filtering_contours(isdebug=isdebug) is False:
return 0.0
#
#https://namkeenman.wordpress.com/2015/12/18/open-cv-determine-angle-of-rotatedrect-minarearect/
#https://docs.opencv.org/2.4/modules/core/doc/basic_structures.html?highlight=rotatedrect#RotatedRect::RotatedRect()
#
# calculate x range of LP
self.estimateLP(isdebug=isdebug)
if isdebug:
tmp = self.bgr.copy()
cv2.line(tmp, (self.charleftmost, self.height / 2),
(self.charrightmost, self.height / 2), (34, 222, 0), 2)
showResult("test", tmp)
# crop and warp, then segment
self.correctImg(isdebug=isdebug)
# detect colr
#self.detectcolr()
# makecompose
self.makeCompose()
# makesegments
self.makesegments(self.compose)
# calculate confidence
self.calculateconfidence()
return self.confidence
def showAll(self):
showResult("img", self.bgr)
#showResult("gray",self.gray)
#showResult("sobel",self.sobel)
#showResult("entropy",self.entropy)
showResult("DoG", self.DoG)
#showResult("garbor",self.garbor)
#showResult("laplacian",self.laplacian)
#showResult("thr",self.lthr)
#showResult("contour",self.contour)
showResult("compose", self.compose)
def detect_by_contour(origin, isdebug=False):
# Initialize
height, width, numChannels = origin.shape
img_gray = np.zeros((height, width, 1), np.uint8)
img_thr = np.zeros((height, width, 1), np.uint8)
img_contour = np.zeros((height, width, 3), np.uint8)
# Grayscale
img_gray, img_thr = preprocess(origin)
if isdebug:
showResult("img_gray", img_gray)
showResult("img_thr", img_thr)
#showResult("Test",cv2.Canny(img_gray,50,200))
# First Filtering(Contours2Chars)
chars = find_possible_chars(img_thr)
if isdebug:
print "step 2 - the numbder of suspicious chars(roughly filtered contours) = " + str(
len(chars))
drawChars(img_contour, "first-filtering", chars, isdebug=True)
# Second Filtering(Chars2Strings)
strings = Contour.findListOfListsOfMatchingChars(chars)
if isdebug: # show steps #######################################################
print "step 3 - strings.Count = " + str(
len(strings)) # 13 with MCLRNF1 image
img_contour = np.zeros((height, width, 3), np.uint8)
for string in strings:
(b, g, r) = (random.randint(0, 255), random.randint(0, 255),
random.randint(0, 255))
img_contour = drawChars(img_contour, "second-filtering", string,
(b, g, r))
# Third Filtering(String2ROIs)
ROIs = []
bboxes = []
for string in strings: # for each group of matching chars
roi = extractROI(origin, string) # attempt to extract plate
if roi.imgPlate is not None: # if plate was found
ROIs.append(roi) # add to list of possible plates
bbox = cv2.boxPoints(roi.rrLocationOfPlateInScene)
ibbox = bbox.astype(int)
bboxes.append(ibbox)
#
print "\n" + str(
len(ROIs)) + " possible plates found" # 13 with MCLRNF1 image
if isdebug and len(
ROIs
) != 0: # show steps #######################################################
for i in range(len(ROIs)):
p2fRectPoints = cv2.boxPoints(ROIs[i].rrLocationOfPlateInScene)
cv2.line(origin, tuple(p2fRectPoints[0]), tuple(p2fRectPoints[1]),
SCALAR_RED, 2)
cv2.line(origin, tuple(p2fRectPoints[1]), tuple(p2fRectPoints[2]),
SCALAR_RED, 2)
cv2.line(origin, tuple(p2fRectPoints[2]), tuple(p2fRectPoints[3]),
SCALAR_RED, 2)
cv2.line(origin, tuple(p2fRectPoints[3]), tuple(p2fRectPoints[0]),
SCALAR_RED, 2)
showResult("roi", ROIs[i].imgPlate)
showResult("result", origin)
return bboxes
def drawChars(self, title, colr=SCALAR_WHITE, isdebug=False):
img_contours = np.zeros((self.bgr.shape), np.uint8)
cv2.drawContours(img_contours, self.contours, -1, colr, -1)
if isdebug:
showResult(title, img_contours)
return img_contours
def debug(self, img, xs):
tmp = img.copy()
for x in xs:
cv2.circle(tmp, (int(x), self.height / 2), 2, (255, 0, 0), 2)
showResult("debug", tmp)
def showAll(self):
showResult("roi", self.bgr)
showResult("gray", self.gray)
showResult("laplacian", self.laplacian)
showResult("DoG", self.DoG)
showResult("thr", self.thr)
def showall(self):
showResult("laplacian", self.laplacian)
showResult("lthr", self.lthr)
#showResult("DoG",self.DoG)
showResult("tophat", self.tophat)
showResult("compose", self.compose)
def labels2boundaries(labels):
#http://scikit-image.org/docs/dev/api/skimage.segmentation.html#skimage.segmentation.find_boundaries
cleared = clear_border(labels)
boundary = find_boundaries(cleared, mode='inner').astype(np.uint8)
boundary[boundary != 0] = 255
showResult("boundaries",boundary)