139 lines
5.5 KiB
Python
139 lines
5.5 KiB
Python
import math
|
|
import cv2
|
|
import numpy
|
|
import matplotlib.pyplot as plt
|
|
|
|
#自动将图像像素值转化成0-1的灰度图像并求简单的图像复原性能指标
|
|
|
|
#image0:滤波前图像 image1:滤波后图像 图像的长或宽不能为3个像素
|
|
|
|
def standardization(img): #将0-255范围的图像转化为0-1，将彩色图像转换为灰度图像
|
|
if len(img.shape)>2 :#判断img.shape元组的长度
|
|
[m, n, k] = img.shape
|
|
if k!=1:
|
|
img = cv2.cvtColor(img, cv2.COLOR_BGR2GRAY)
|
|
else : #img.shape长度为2，则是灰度图像
|
|
[m ,n] = img.shape
|
|
for i in range(m):
|
|
for j in range(n):
|
|
if(img[i,j]!=0):
|
|
if(img[i,j]>1):
|
|
img = img / 255
|
|
break
|
|
return img
|
|
|
|
def meanSquare(image0,image1):#求均方误差
|
|
image0 = standardization(image0)
|
|
image1 = standardization(image1)
|
|
[m, n] = image0.shape
|
|
MSE=0
|
|
for i in range(m):
|
|
for j in range(n):
|
|
MSE=MSE+(image0[i,j]-image1[i,j])**2
|
|
MSE=MSE/(m*n)
|
|
return MSE
|
|
|
|
def SignalNoiseRatio(image0,image1):#求信噪比，输出单位：dB
|
|
image0 = standardization(image0)
|
|
image1 = standardization(image1)
|
|
[m, n] = image0.shape
|
|
Np=0
|
|
Sp=0
|
|
for i in range(m):
|
|
for j in range(n):
|
|
Np=Np+(image0[i,j]-image1[i,j])**2
|
|
Sp=Sp+image1[i,j]**2
|
|
SNR=10*math.log(Sp/(Np+1e-14),10) #防止除数为0
|
|
return SNR
|
|
|
|
def PeakSignalNoiseRatio(image0,image1):#求峰值信噪比，输出单位：dB
|
|
image0 = standardization(image0)
|
|
image1 = standardization(image1)
|
|
[m, n] = image0.shape
|
|
MSE=0
|
|
for i in range(m):
|
|
for j in range(n):
|
|
MSE = MSE + (image0[i, j] - image1[i, j]) ** 2
|
|
MSE = MSE / (m * n)
|
|
PSNR=10*math.log(1/(MSE+1e-14),10) #防止除数为0
|
|
return PSNR
|
|
|
|
def SimpleRPMPrint(image0,image1):#输出均方误差，信噪比，峰值信噪比
|
|
MSE=meanSquare(image0,image1)
|
|
SNR=SignalNoiseRatio(image0,image1)
|
|
PSNR=PeakSignalNoiseRatio(image0,image1)
|
|
print("均方误差为：",MSE)
|
|
print("信噪比为：", SNR)
|
|
print("峰值信噪比为：", PSNR)
|
|
return
|
|
|
|
def SignalNoiseRatio(image0,image1):#求信噪比，输出单位：dB
|
|
image0 = standardization(image0)
|
|
image1 = standardization(image1)
|
|
[m, n] = image0.shape
|
|
Np=0
|
|
Sp=0
|
|
for i in range(m):
|
|
for j in range(n):
|
|
Np=Np+(image0[i,j]-image1[i,j])**2
|
|
Sp=Sp+image1[i,j]**2
|
|
SNR=10*math.log(Sp/(Np+1e-14),10) #防止除数为0
|
|
return SNR
|
|
def PeakSignalNoiseRatio(image0,image1):#求峰值信噪比，输出单位：dB
|
|
image0 = standardization(image0)
|
|
image1 = standardization(image1)
|
|
[m, n] = image0.shape
|
|
MSE=0
|
|
for i in range(m):
|
|
for j in range(n):
|
|
MSE = MSE + (image0[i, j] - image1[i, j]) ** 2
|
|
MSE = MSE / (m * n)
|
|
PSNR=10*math.log(1/(MSE+1e-14),10) #防止除数为0
|
|
return PSNR
|
|
# #以下是信噪比曲线，lp0理想低通，lp1巴特沃斯低通，lp2高斯低通，hp0理想高通，hp1巴特沃斯高通，hp2高斯高通
|
|
# a1 = [0] * 80
|
|
# a2 = [0] * 80
|
|
# a3 = [0] * 80
|
|
# a4 = [0] * 80
|
|
# a5 = [0] * 80
|
|
# a6 = [0] * 80
|
|
# img1 = cv2.imread('c:/9.jpg', 0)
|
|
# for i in range(20, 100):
|
|
# a1[i - 20] = SignalNoiseRatio(img1, ifft(lpfilter(0, 375, 375, i, 1) * fft_mat))
|
|
# a2[i - 20] = SignalNoiseRatio(img1, ifft(lpfilter(1, 375, 375, i, 1) * fft_mat))
|
|
# a3[i - 20] = SignalNoiseRatio(img1, ifft(lpfilter(2, 375, 375, i, 1) * fft_mat))
|
|
# a4[i - 20] = SignalNoiseRatio(img1, ifft(hpfilter(0, 375, 375, i, 1) * fft_mat))
|
|
# a5[i - 20] = SignalNoiseRatio(img1, ifft(hpfilter(1, 375, 375, i, 1) * fft_mat))
|
|
# a6[i - 20] = SignalNoiseRatio(img1, ifft(hpfilter(2, 375, 375, i, 1) * fft_mat))
|
|
# x = np.linspace(20, 80, 80)
|
|
# plt.figure(figsize=(10, 7))
|
|
# plt.plot(x, a1, lw=2.0, ls='-', color='r', label='lp_0')
|
|
# plt.plot(x, a2, lw=2.0, ls='-', color='b', label='lp_1')
|
|
# plt.plot(x, a3, lw=2.0, ls='-', color='y', label='lp_2')
|
|
# plt.plot(x, a4, lw=2.0, ls='-', color='k', label='hp_0')
|
|
# plt.plot(x, a5, lw=2.0, ls='-', color='c', label='hp_1')
|
|
# plt.plot(x, a6, lw=2.0, ls='-', color='g', label='hp_2')
|
|
# plt.xlabel("d0", fontsize=12)
|
|
# plt.ylabel("SNR", fontsize=12)
|
|
# plt.legend(loc='upper right', fontsize=12)
|
|
# plt.show()
|
|
|
|
if __name__=='__main__':
|
|
img = cv2.imread(r'.\img\lena.bmp',0)
|
|
dft = cv2.dft(np.float32(img), flags=cv2.DFT_COMPLEX_OUTPUT) # 对图像进行傅里叶变换
|
|
# dft[:,:,0]为傅里叶变换的实部,dft[:,:,1]为傅里叶变换的虚部
|
|
magnitude0 = 20 * np.log(1 + cv2.magnitude(dft[:, :, 0], dft[:, :, 1])) # 幅值谱
|
|
phase0 = cv2.phase(dft[:, :, 0], dft[:, :, 1]) # 相位谱
|
|
dft_shift = np.fft.fftshift(dft)
|
|
magnitude1 = 20 * np.log(1 + cv2.magnitude(dft_shift[:, :, 0], dft_shift[:, :, 1])) # 幅值谱
|
|
phase1 = cv2.phase(dft_shift[:, :, 0], dft_shift[:, :, 1]) # 相位谱
|
|
plt.subplot(221),plt.imshow(img, cmap = 'gray')
|
|
plt.title('Input Image'), plt.xticks([]), plt.yticks([])
|
|
plt.subplot(222),plt.imshow(phase1, cmap = 'gray')
|
|
plt.title('phase angle'), plt.xticks([]), plt.yticks([])
|
|
plt.subplot(223),plt.imshow(magnitude0, cmap = 'gray')
|
|
plt.title('magnitude spectrum'), plt.xticks([]), plt.yticks([])
|
|
plt.subplot(224),plt.imshow(magnitude1, cmap = 'gray') #显示移中后的幅值谱
|
|
plt.title('magnitude_spectrum after shift'), plt.xticks([]), plt.yticks([])
|
|
plt.show()
|