-
Notifications
You must be signed in to change notification settings - Fork 0
/
background.py
182 lines (168 loc) · 6.68 KB
/
background.py
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
93
94
95
96
97
98
99
100
101
102
103
104
105
106
107
108
109
110
111
112
113
114
115
116
117
118
119
120
121
122
123
124
125
126
127
128
129
130
131
132
133
134
135
136
137
138
139
140
141
142
143
144
145
146
147
148
149
150
151
152
153
154
155
156
157
158
159
160
161
162
163
164
165
166
167
168
169
170
171
172
173
174
175
176
177
178
179
180
181
182
# -*- coding: utf-8 -*-
import numpy as np
import matplotlib.pyplot as plt
import cv2
from math import log
from PIL import Image
import datetime
import pywt
# 以下强行用Python宏定义变量
halfWindowSize=9
src1_path = '2.jpg'
src2_path = '203.jpg'
'''''
来自敬忠良，肖刚，李振华《图像融合——理论与分析》P85：基于像素清晰度的融合规则
1，用Laplace金字塔或者是小波变换，将图像分解成高频部分和低频部分两个图像矩阵
2，以某个像素点为中心开窗，该像素点的清晰度定义为窗口所有点((高频/低频)**2).sum()
3，目前感觉主要的问题在于低频
4，高频取清晰度图像中较大的那个图的高频图像像素点
5，算法优化后速度由原来的2min.44s.变成9s.305ms.
补充：书上建议开窗大小10*10，DWT取3层，Laplace金字塔取2层
'''
def imgOpen(img_src1,img_src2):
apple=Image.open(img_src1).convert('L')
orange=Image.open(img_src2).convert('L')
appleArray=np.array(apple)
orangeArray=np.array(orange)
return appleArray,orangeArray
# 严格的变换尺寸
def _sameSize(img_std,img_cvt):
x,y=img_std.shape
pic_cvt=Image.fromarray(img_cvt)
pic_cvt.resize((x,y))
return np.array(pic_cvt)
# 小波变换的层数不能太高，Image模块的resize不能变换太小的矩阵，不相同大小的矩阵在计算对比度时会数组越界
def getWaveImg(apple,orange):
appleWave=pywt.wavedec2(apple,'haar',level=4)
orangeWave=pywt.wavedec2(orange,'haar',level=4)
lowApple=appleWave[0];lowOrange=orangeWave[0]
# 以下处理低频
lowAppleWeight,lowOrangeWeight = getVarianceWeight(lowApple,lowOrange)
lowFusion = lowAppleWeight*lowApple + lowOrangeWeight*lowOrange
# 以下处理高频
for hi in range(1,5):
waveRec=[]
for highApple,highOrange in zip(appleWave[hi],orangeWave[hi]):
highFusion = np.zeros(highApple.shape)
contrastApple = getContrastImg(lowApple,highApple)
contrastOrange = getContrastImg(lowOrange,highOrange)
row,col = highApple.shape
for i in xrange(row):
for j in xrange(col):
if contrastApple[i,j] > contrastOrange[i,j]:
highFusion[i,j] = highApple[i,j]
else:
highFusion[i,j] = highOrange[i,j]
waveRec.append(highFusion)
recwave=(lowFusion,tuple(waveRec))
lowFusion=pywt.idwt2(recwave,'haar')
lowApple=lowFusion;lowOrange=lowFusion
return lowFusion
# 求Laplace金字塔
def getLaplacePyr(img):
firstLevel=img.copy()
secondLevel=cv2.pyrDown(firstLevel)
lowFreq=cv2.pyrUp(secondLevel)
highFreq=cv2.subtract(firstLevel,_sameSize(firstLevel,lowFreq))
return lowFreq,highFreq
# 计算对比度，优化后不需要这个函数了，扔在这里看看公式就行
def _getContrastValue(highWin,lowWin):
row,col = highWin.shape
contrastValue = 0.00
for i in xrange(row):
for j in xrange(col):
contrastValue += (float(highWin[i,j])/lowWin[i,j])**2
return contrastValue
# 先求出每个点的(hi/lo)**2，再用numpy的sum（C语言库）求和
def getContrastImg(low,high):
row,col=low.shape
if low.shape!=high.shape:
low=_sameSize(high,low)
contrastImg=np.zeros((row,col))
contrastVal=(high/low)**2
for i in xrange(row):
for j in xrange(col):
up=i-halfWindowSize if i-halfWindowSize>0 else 0
down=i+halfWindowSize if i+halfWindowSize<row else row
left=j-halfWindowSize if j-halfWindowSize>0 else 0
right=j+halfWindowSize if j+halfWindowSize<col else col
contrastWindow=contrastVal[up:down,left:right]
contrastImg[i,j]=contrastWindow.sum()
return contrastImg
# 计算方差权重比
def getVarianceWeight(apple,orange):
appleMean,appleVar=cv2.meanStdDev(apple)
orangeMean,orangeVar=cv2.meanStdDev(orange)
appleWeight=float(appleVar)/(appleVar+orangeVar)
orangeWeight=float(orangeVar)/(appleVar+orangeVar)
return appleWeight,orangeWeight
# 函数返回融合后的图像矩阵
def getPyrFusion(apple,orange):
lowApple,highApple = getLaplacePyr(apple)
lowOrange,highOrange = getLaplacePyr(orange)
contrastApple = getContrastImg(lowApple,highApple)
contrastOrange = getContrastImg(lowOrange,highOrange)
row,col = lowApple.shape
highFusion = np.zeros((row,col))
lowFusion = np.zeros((row,col))
# 开始处理低频
# appleWeight,orangeWeight=getVarianceWeight(lowApple,lowOrange)
for i in xrange(row):
for j in xrange(col):
# lowFusion[i,j]=lowApple[i,j]*appleWeight+lowOrange[i,j]*orangeWeight
lowFusion[i,j] = lowApple[i,j] if lowApple[i,j]<lowOrange[i,j] else lowOrange[i,j]
# 开始处理高频
for i in xrange(row):
for j in xrange(col):
highFusion[i,j] = highApple[i,j] if contrastApple[i,j] > contrastOrange[i,j] else highOrange[i,j]
# 开始重建
fusionResult = cv2.add(highFusion,lowFusion)
return fusionResult
# 绘图函数
def getPlot(apple,orange,result):
plt.subplot(131)
plt.imshow(apple,cmap='gray')
plt.title('src1')
plt.axis('off')
plt.subplot(132)
plt.imshow(orange,cmap='gray')
plt.title('src2')
plt.axis('off')
plt.subplot(133)
plt.imshow(result,cmap='gray')
plt.title('result')
plt.axis('off')
plt.show()
# 画四张图的函数，为了方便同时比较
def cmpPlot(apple,orange,wave,pyr):
plt.subplot(221)
plt.imshow(apple,cmap='gray')
plt.title('SRC1')
plt.axis('off')
plt.subplot(222)
plt.imshow(orange,cmap='gray')
plt.title('SRC2')
plt.axis('off')
plt.subplot(223)
plt.imshow(wave,cmap='gray')
plt.title('WAVELET')
plt.axis('off')
plt.subplot(224)
plt.imshow(pyr,cmap='gray')
plt.title('LAPLACE PYR')
plt.axis('off')
plt.show()
def runTest(src1=src1_path,src2=src2_path,isplot=True):
apple,orange=imgOpen(src1,src2)
beginTime=datetime.datetime.now()
print(beginTime)
waveResult=getWaveImg(apple,orange)
pyrResult=getPyrFusion(apple,orange)
endTime=datetime.datetime.now()
print(endTime)
print('Runtime: '+str(endTime-beginTime))
if isplot:
cmpPlot(apple,orange,waveResult,pyrResult)
return waveResult,pyrResult
if __name__=='__main__':
runTest()