-
Notifications
You must be signed in to change notification settings - Fork 10
/
Copy path图片比对工具python版.py
247 lines (221 loc) · 9.25 KB
/
图片比对工具python版.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
183
184
185
186
187
188
189
190
191
192
193
194
195
196
197
198
199
200
201
202
203
204
205
206
207
208
209
210
211
212
213
214
215
216
217
218
219
220
221
222
223
224
225
226
227
228
229
230
231
232
233
234
235
236
237
238
239
240
241
242
243
244
245
246
247
#coding:utf8
import sys
import os
import cv2
import numpy as np
def AssertImage( old_img, new_img, diff_img,check=None, nocheck=None,standard=90):
'''
old_img:原始的图片
new_img:需要对比的图片
diff_img:输出显示比对区域的图片
standard:相似度结果合格的标准数值，默认为90，可由前台设置
check：需要对比的区域,(左上角坐标，又下角坐标)
check = [
((0,0),(0,1)),
((10,15),(50,70)),
]
nocheck：不需要对比的区域
nocheck = [
((0,0),(0,1)),
((10,15),(50,70)),
]
当check和nocheck同时为空的时候，对比整张图片
'''
res = []
deep = 3 # 检测深度，作为每块长宽的分母，deep值越大，分的块越小，检测越严格
result = 0
bResult = 100
'''
if check is 'None':
check = []
else:
#将字符串str当成有效的表达式来求值并返回计算结果
check = eval(check)
if nocheck is 'None':
nocheck = []
else:
#将字符串str当成有效的表达式来求值并返回计算结果
nocheck = eval(nocheck)
'''
if not os.path.exists(old_img):
err = "Image({0}) not exeist".format(old_img)
print (err)
raise RunActivceException(err)
if not os.path.exists(new_img):
err = "Image({0}) not exeist".format(new_img)
print (err)
raise RunActivceException(err)
#cv2.imread() 读取图片，并显示
cv2_old_img = cv2.imread(old_img)
cv2_new_img = cv2.imread(new_img)
#图像矩阵的shape属性表示图像的大小，shape会返回tuple元组，第一个元素表示矩阵行数，第二个元组表示矩阵列数，第三个元素是3，表示像素值由光的三原色组成
oh, ow, _ = cv2_old_img.shape
nh, nw, _ = cv2_new_img.shape
#图片不一样大
if oh != nh or ow != nw:
print("The resolution of the old img is:[{0}*{1}]".format(oh, ow))
print("The resolution of the new img is:[{0}*{1}]".format(nh, nw))
raise RunActivceException("The resolution of the two pictures is different")
# 此处保留原本算法，用于计算整体相似度，并给出比对区域的图片
if nocheck is None:
nocheck = []
if check is None:
compare(cv2_new_img, cv2_old_img, check, nocheck, diff_img=diff_img,oh=oh, ow=ow, nh=nh, nw=nw)
else:
for ((sx, sy), (ex, ey)) in check:
compare(cv2_new_img, cv2_old_img, check, nocheck, diff_img=diff_img,
sx=sx, sy=sy, ex=ex, ey=ey)
# 此处开始精确比对，深度deep默认为3
oh /= deep # 图片块的宽度
ow /= deep
nh /= deep
nw /= deep
output = None
if check is None:
for t in range(deep):
for s in range(deep):
bResult, output = compare(cv2_new_img, cv2_old_img, check, nocheck,
t, s, oh=oh, ow=ow, nh=nh, nw=nw,standard=standard)
res.append(100 - bResult)
else:
for ((sx, sy), (ex, ey)) in check:
a = (ex - sx) / deep # 图片块的长度
b = (ey - sy) / deep # 图片块的宽度
for t in range(deep):
for s in range(deep):
bResult, output = compare(cv2_new_img, cv2_old_img, check, nocheck,
t, s, sx=a * s + sx, sy=b * t + sy, ex=a * (s + 1) + sx,
ey=b * (t + 1) + sy,standard=standard)
res.append(100 - bResult)
res.sort()
final_result = sum(res[:3]) / len(res[:3])
output = 'The image similarity is [{0}%],standard is [{1}%]'.format(final_result, standard)
print (output)
return final_result, [output, old_img, new_img, diff_img]
def compare( cv2_new_img, cv2_old_img, check, nocheck, t=0, s=0, diff_img=None, oh=0, ow=0, nh=0, nw=0, sx=0,
sy=0, ex=0, ey=0,standard=90):
bResult = 100
result = 0
if oh != nh or ow != nw:
print ("The resolution of the old img is:[{0}*{1}]".format(oh, ow))
print ("The resolution of the new img is:[{0}*{1}]".format(nh, nw))
print ("The resolution of the two pictures is different")
#empty 会创建一个没有使用特定值来初始化的数组
cv2_diff_img = np.empty(cv2_old_img.shape, np.uint8)
#初始化为255
cv2_diff_img.fill(255)
## 权重越大，透明度越低 标量加到每一个和 函数把两个图片融合在一起
cv2_old_img = cv2.addWeighted(cv2_old_img, 0.4, cv2_diff_img, 0.5, 0)
cv2_new_img = cv2.addWeighted(cv2_new_img, 0.4, cv2_diff_img, 0.5, 0)
cv2_diff_img = cv2_old_img.copy()
#宽
width = cv2_old_img.shape[0]
#高
height = cv2_old_img.shape[1]
if check is not None:
tmp = 0
# sx/=2
# ex/=2
continueFlag = False
#x 方向的像素
for i in range(sx, ex):
if continueFlag:
continueFlag = False
continue
#y方向的像素
for j in range(sy, ey):
for ((nosx, nosy), (noex, noey)) in nocheck:
#不在对比范围之内
if (nosx < i and i < noex) and (nosy < j and j < noey):
# print "check point:({0}) in no check point {1}-{2}".format((i,j),(nosx,nosy),(noex,noey))
continueFlag = True
break
if continueFlag:
continue
diff = int(cv2_old_img[j, i][0]) - int(cv2_new_img[j, i][0])
tmp += 1
if diff < 0:
if _clmap(abs(diff), 10, 255) != 255:
result += 1
cv2_diff_img[j, i] = [0, _clmap(abs(diff), 10, 255), 0]
elif diff > 0:
if _clmap(abs(diff), 10, 255) != 255:
result += 1
cv2_diff_img[j, i] = [0, 0, _clmap(abs(diff), 10, 255)]
else:
pass
if tmp != 0:
bResult = float(result * 100) / float(tmp)
#v2.rectangle(img, (x,y), (x+w,y+h), (0,255,0), 2)画出矩行
#参数解释
#第一个参数：img是原图
#第二个参数：（x，y）是矩阵的左上点坐标
#第三个参数：（x+w，y+h）是矩阵的右下点坐标
#第四个参数：（0,255,0）是画线对应的rgb颜色
#第五个参数：2是所画的线的宽度
cv2.rectangle(cv2_diff_img, (sx, sy), (ex, ey), (0, 255, 0), thickness=3, lineType=16)
else:
for i in range(oh * t, oh * (t + 1)):
continueFlag = False
for j in range(ow * s, ow * (s + 1)):
if continueFlag:
continueFlag = False
continue
for ((nosx, nosy), (noex, noey)) in nocheck:
if (nosx < j and j < noex) and (nosy < i and i < noey):
# print "check point:({0}) in no check point {1}-{2}".format((i,j),(nosx,nosy),(noex,noey))
continueFlag = True
break
if continueFlag:
continue
diff = int(cv2_old_img[i, j][0]) - int(cv2_new_img[i, j][0])
if diff < 0:
if _clmap(abs(diff), 10, 255) != 255:
result += 1
cv2_diff_img[i, j] = [0, _clmap(abs(diff), 10, 255), 0]
elif diff > 0:
if _clmap(abs(diff), 10, 255) != 255:
result += 1
cv2_diff_img[i, j] = [0, 0, _clmap(abs(diff), 10, 255)]
else:
pass
# result +=(abs(diff))
bResult = float(result * 100) / float(oh * ow)
cv2.rectangle(cv2_diff_img, (ow * s, oh * t), (ow * (s + 1), oh * (t + 1)), (0, 255, 0),
thickness=3, lineType=16)
for ((nosx, nosy), (noex, noey)) in nocheck:
cv2.rectangle(cv2_diff_img, (nosx, nosy), (noex, noey), (255, 0, 0), thickness=3, lineType=16)
if diff_img is not None:
output = 'The image general similarity is [{0}%],standard is [{1}%]'.format(100 - bResult,
standard)
cv2.imwrite(diff_img, cv2_diff_img)
else:
output = 'The image area({0},{1}) similarity is [{2}%],standard is [{3}%]'.format(t, s, 100 - bResult,
standard)
print (output)
return bResult, output
def _clmap( v, k, upBound):
val = v * k
if standard == 'None':
standard = int(90)
else:
standard = int(standard)
if val > upBound:
return val
else:
return upBound
if __name__ == '__main__':
args = sys.argv[:]
print args
old_img = sys.argv[1]
new_img = sys.argv[2]
diff_img = sys.argv[3]
check = sys.argv[4]
nocheck = sys.argv[5]
standard = sys.argv[6]
print type(check)
check = eval(check)
nocheck = eval(nocheck)
standard = eval(standard)
AssertImage(old_img,new_img,diff_img, check=check, nocheck=nocheck, standard=standard)
#AssertImage('112.png','111.png','diff.png')