def update(val):
H = myFunc.getBWlp(P, Q, int(sn.val), int(sD0.val))
temp = myFFT.ifft2(myFFT.fftshift(np.multiply(F , H)),p,q)
#temp = unpad2(temp, P, Q)
temp = np.uint8(np.abs(temp))
self.onPanel2(temp)
fig.add_subplot(223)
plt.imshow(temp, cmap = plt.cm.gray)
plt.yticks([])
plt.xticks([])
plt.title("Filtered Image")
fig.add_subplot(224)
plt.imshow(H, cmap = plt.cm.gist_heat)
plt.yticks([])
plt.xticks([])
plt.title("Frequency Response")
def onVwSpec(self):
self.I2 = self.Ilast
if self.Ilast.ndim == 3 :
tkMessageBox.showinfo("Message", "Image will be converted to Grayscale.")
self.I2 = myThresh.gray(self.I2)
# padding before taking Fourier Transoffrm
#temp, _, _ = pad2(self.I2)
#temp = self.I2
temp = np.float64(self.I2)
F,_,_ = myFFT.fft2(temp)
F = myFFT.fftshift(F)
pwrSpec = np.log10(1 + np.abs(F) ** 2)
plt.imshow(pwrSpec, cmap = plt.cm.PRGn)
plt.yticks([])
plt.xticks([])
plt.title("Power Spectrum")
cbar = plt.colorbar(orientation='horizontal')
cbar.set_label('Frequency Amplitudes')
plt.show()
def update(val):
H = 1.0 - myFunc.getBWlp(P, Q, int(sn.val), int(sD0.val))
temp = myFFT.ifft2(myFFT.fftshift(np.multiply(F , H)),p,q)
temp = self.I2 + temp#unpad2(temp, P, Q)
np.putmask(temp, temp > 255, 255) # check overflow
np.putmask(temp, temp < 0, 0) # check underflow
temp = np.uint8(np.abs(temp))
self.onPanel2(temp)
fig.add_subplot(223)
plt.imshow(temp, cmap = plt.cm.gray)
plt.yticks([])
plt.xticks([])
plt.title("Filtered Image")
fig.add_subplot(224)
plt.imshow(H, cmap = plt.cm.gist_heat)
plt.yticks([])
plt.xticks([])
plt.title("Frequency Response")
def onBwLp(self):
self.I2 = self.Ilast
if self.Ilast.ndim == 3 :
tkMessageBox.showinfo("Message", "Image will be converted to Grayscale.")
self.I2 = myThresh.gray(self.I2)
#tempI, P, Q = pad2(self.I2)
tempI = np.float64(self.I2)
F,p,q = myFFT.fft2(tempI)
P, Q = F.shape
F = myFFT.fftshift(F)
n = 10
D0 = np.min([P,Q])/4
H = myFunc.getBWlp(P, Q, n, D0)
temp = myFFT.ifft2(myFFT.fftshift(np.multiply(F , H)),p,q)
#temp = unpad2(temp, P, Q)
temp = np.uint8(np.abs(temp))
self.onPanel2(temp)
plt.clf()
fig = plt.figure(figsize=(16,9),facecolor='w')
fig.add_subplot(221)
plt.subplots_adjust(left = 0, bottom=0.06, right = 1, top = 0.95, hspace = 0.1, wspace = 0)
plt.imshow(self.I2, cmap = plt.cm.gray)
plt.yticks([])
plt.xticks([])
plt.title("Original Image")
fig.add_subplot(222)
pwrSpec = np.log10(1 + np.abs(F) ** 2)
plt.imshow(pwrSpec, cmap = plt.cm.PRGn)
plt.yticks([])
plt.xticks([])
plt.title("Power Spectrum")
cbar = plt.colorbar(ticks = [])
cbar.set_label(r'Spectral Power')
fig.add_subplot(223)
plt.imshow(temp, cmap = plt.cm.gray)
plt.yticks([])
plt.xticks([])
plt.title("Filtered Image")
fig.add_subplot(224)
plt.imshow(H, cmap = plt.cm.gist_heat)
plt.yticks([])
plt.xticks([])
plt.title("Frequency Response")
cbar = plt.colorbar(ticks = [])
cbar.set_label(r'Amplitude')
axcolor = 'lightgoldenrodyellow'
axD0 = fig.add_axes([0.1, 0.025, 0.3, 0.025], axisbg=axcolor)
axn = fig.add_axes([0.6, 0.025, 0.3, 0.025], axisbg=axcolor)
sD0 = Slider(axD0, 'Cut Off freq.', 20, np.min([P / 40 * 20, Q / 40 * 20]), valinit=D0)
sn = Slider(axn, "degree 'n'", 1, 25, valinit=n)
def update(val):
H = myFunc.getBWlp(P, Q, int(sn.val), int(sD0.val))
temp = myFFT.ifft2(myFFT.fftshift(np.multiply(F , H)),p,q)
#temp = unpad2(temp, P, Q)
temp = np.uint8(np.abs(temp))
self.onPanel2(temp)
fig.add_subplot(223)
plt.imshow(temp, cmap = plt.cm.gray)
plt.yticks([])
plt.xticks([])
plt.title("Filtered Image")
fig.add_subplot(224)
plt.imshow(H, cmap = plt.cm.gist_heat)
plt.yticks([])
plt.xticks([])
plt.title("Frequency Response")
sD0.on_changed(update)
sn.on_changed(update)
fig.show()
def onGaussianHp(self):
self.I2 = self.Ilast
if self.Ilast.ndim == 3 :
tkMessageBox.showinfo("Message", "Image will be converted to Grayscale.")
self.I2 = myThresh.gray(self.I2)
#tempI, P, Q = pad2(self.I2)
tempI = np.float64(self.I2)
F,p,q = myFFT.fft2(tempI)
P, Q = F.shape
F = myFFT.fftshift(F)
sigma = np.min([P,Q]) / 5
H = 1.0 - myFunc.getGaussianlp(P, Q, sigma)
temp = myFFT.ifft2(myFFT.fftshift(np.multiply(F , H)),p,q)
temp = self.I2 + temp
np.putmask(temp, temp > 255, 255) # check overflow
np.putmask(temp, temp < 0, 0) # check underflow
temp = np.uint8(np.abs(temp))
self.onPanel2(temp)
plt.clf()
fig = plt.figure(figsize=(16,9),facecolor='w')
fig.add_subplot(221)
plt.subplots_adjust(left = 0, bottom=0.06, right = 1, top = 0.95, hspace = 0.1, wspace = 0)
plt.imshow(self.I2, cmap = plt.cm.gray)
plt.yticks([])
plt.xticks([])
plt.title("Original Image")
fig.add_subplot(222)
pwrSpec = np.log10(1 + np.abs(F) ** 2)
plt.imshow(pwrSpec, cmap = plt.cm.PRGn)
plt.yticks([])
plt.xticks([])
plt.title("Power Spectrum")
cbar = plt.colorbar(ticks = [])
cbar.set_label(r'Spectral Power')
fig.add_subplot(223)
plt.imshow(temp, cmap = plt.cm.gray)
plt.yticks([])
plt.xticks([])
plt.title("Filtered Image")
fig.add_subplot(224)
plt.imshow(H, cmap = plt.cm.gist_heat)
plt.yticks([])
plt.xticks([])
plt.title("Frequency Response")
cbar = plt.colorbar(ticks = [])
cbar.set_label(r'Amplitude')
axcolor = 'lightgoldenrodyellow'
axSigma = fig.add_axes([0.2, 0.025, 0.5, 0.025], axisbg=axcolor)
sSigma = Slider(axSigma, 'Cut Off freq.( Sigma)', 5, np.min([P / 50 * 20, Q / 50 * 20]), valinit=sigma)
def update(val):
H = 1.0 - myFunc.getGaussianlp(P, Q, int(sSigma.val))
temp = myFFT.ifft2(myFFT.fftshift(np.multiply(F , H)), p, q)
temp = self.I2 + temp
np.putmask(temp, temp > 255, 255) # check overflow
np.putmask(temp, temp < 0, 0) # check underflow
temp = np.uint8(np.abs(temp))
self.onPanel2(temp)
fig.add_subplot(223)
plt.imshow(temp, cmap = plt.cm.gray)
plt.yticks([])
plt.xticks([])
plt.title("Filtered Image")
fig.add_subplot(224)
plt.imshow(H, cmap = plt.cm.gist_heat)
plt.yticks([])
plt.xticks([])
plt.title("Frequency Response")
sSigma.on_changed(update)
fig.show()