def MatchAll(I, template, thr):
Ibcp = copy(I)
(w,h,t) = shape(template)
CrossCorr = cv2.matchTemplate(I, template, cv2.TM_CCORR_NORMED)
cv2.normalize(CrossCorr,CrossCorr,0,255,cv2.NORM_MINMAX)
for idxr, r in enumerate(CrossCorr):
for idxc, c in enumerate(r):
if c>= thr:
x=idxc
y=idxr
x1=idxc+h
x2=idxr+w
cv2.rectangle(Ibcp, (x,y), (x1,x2),(60,45,255),2)
tools.show_Multiple_images_pylab(Oryginal=I, CrossCor=CrossCorr, result = Ibcp)
def MatchAll(I, template, thr):
Ibcp = copy(I)
(w, h, t) = shape(template)
CrossCorr = cv2.matchTemplate(I, template, cv2.TM_CCORR_NORMED)
cv2.normalize(CrossCorr, CrossCorr, 0, 255, cv2.NORM_MINMAX)
for idxr, r in enumerate(CrossCorr):
for idxc, c in enumerate(r):
if c >= thr:
x = idxc
y = idxr
x1 = idxc + h
x2 = idxr + w
cv2.rectangle(Ibcp, (x, y), (x1, x2), (60, 45, 255), 2)
tools.show_Multiple_images_pylab(Oryginal=I,
CrossCor=CrossCorr,
result=Ibcp)
def addGaussianRandomNoise(I, mu, sigma):
M,N=I.shape
gaussianNoise = np.random.normal(mu, sigma,(M,N))
gaussianNoise=tools.to_uint8(gaussianNoise)
I += gaussianNoise
return I
def addRandomUniformNoise(I):
M,N=I.shape
noise = np.random.rand(M,N)*60
noise=tools.to_uint8(noise)#
I += noise
return I
M,N=I.shape
I[((np.random.rand(M,N))<density)] = [0]
I[((np.random.rand(M,N))<density)] = [255]
return I
#------------------------------BODY---------------------------------------
I = cv2.imread("lena.png")
I = cv2.cvtColor(I, cv2.COLOR_RGB2GRAY)
#Dst = BoxFilter(I,7)
#tools.show_Multiple_images_OpenCV(Oryginal=I, Filtered = Dst, Difference = cv2.absdiff(I, Dst))
#tools.show_1_image_OpenCV(addSaltNPepperNoise(I, 0.07))
Iun = addRandomUniformNoise(copy(I))
Igrn = addGaussianRandomNoise(copy(I), 0, 50)
Isp = addSaltNPepperNoise(copy(I), 0.07)
tools.show_Multiple_images_OpenCV(Iun=BoxFilter(Iun, 5), Igrn=BoxFilter(Igrn, 5), SnP=BoxFilter(I, 5))
def addGaussianRandomNoise(I, mu, sigma):
M, N = I.shape
gaussianNoise = np.random.normal(mu, sigma, (M, N))
gaussianNoise = tools.to_uint8(gaussianNoise)
I += gaussianNoise
return I
def addRandomUniformNoise(I):
M, N = I.shape
noise = np.random.rand(M, N) * 60
noise = tools.to_uint8(noise) #
I += noise
return I
gaussianNoise = np.random.normal(mu, sigma, (M, N))
gaussianNoise = tools.to_uint8(gaussianNoise)
I += gaussianNoise
return I
def addSaltNPepperNoise(I, density):
M, N = I.shape
I[((np.random.rand(M, N)) < density)] = [0]
I[((np.random.rand(M, N)) < density)] = [255]
return I
#------------------------------BODY---------------------------------------
I = cv2.imread("lena.png")
I = cv2.cvtColor(I, cv2.COLOR_RGB2GRAY)
#Dst = BoxFilter(I,7)
#tools.show_Multiple_images_OpenCV(Oryginal=I, Filtered = Dst, Difference = cv2.absdiff(I, Dst))
#tools.show_1_image_OpenCV(addSaltNPepperNoise(I, 0.07))
Iun = addRandomUniformNoise(copy(I))
Igrn = addGaussianRandomNoise(copy(I), 0, 50)
Isp = addSaltNPepperNoise(copy(I), 0.07)
tools.show_Multiple_images_OpenCV(Iun=BoxFilter(Iun, 5),
Igrn=BoxFilter(Igrn, 5),
SnP=BoxFilter(I, 5))