-
Notifications
You must be signed in to change notification settings - Fork 0
Expand file tree
/
Copy pathimage_processing.py
More file actions
241 lines (198 loc) · 8.66 KB
/
image_processing.py
File metadata and controls
241 lines (198 loc) · 8.66 KB
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
import cv2,os
from PIL import Image
from os.path import join
import numpy as np
DataSet_Path = "WEBLMT_dataset/"
ClearLine_Path = "WEBLMT_clearline/"
ClearBackColor_Path = "WEBLMT_clearbackcolor/"
ClearNoise_Path = "WEBLMT_clearnoise/"
TrainSet_Path = "WEBLMT_train/"
TestSet_Path = "WEBLMT_test/"
Train_Divide_Path = "WEBLMT_train_divide/"
Test_Divide_Path = "WEBLMT_test_divide/"
Buffer_Path = "WEBLMT_buffer/"
label_filename = "result.txt"
suffix = '.png'
IMAGE_HEIGHT = 16
IMAGE_WEIGHT = 64
def clear_border(img,ratio):
h, w = img.shape[:2]
ROI_img = img[int(h*ratio[1]):int(h*(1-ratio[1])),int(w*ratio[0]):int(w*(1-ratio[0]))]
return ROI_img
def get_dynamic_binary_image(img):
img = cv2.cvtColor(img,cv2.COLOR_BGR2GRAY) #灰值化
# 二值化
th1 = cv2.adaptiveThreshold(img, 255, cv2.ADAPTIVE_THRESH_GAUSSIAN_C, cv2.THRESH_BINARY, 11, 2)
return th1
def get_line_position(line_img,backgroud_value = 255):
lines = []
h, w = line_img.shape[:2]
for x in range(0, h):
for y in range(0,w):
if(line_img[x, y] != backgroud_value):
lines.append(x)
break
return lines
def clear_line(src_img,lines):
h,w = src_img.shape[:2]
for x in lines:
for y in range(w):
src_img[x,y]=np.ceil((src_img[x-1,y]/2+src_img[x+1,y]/2))
return src_img
def clear_line_auto(src):
gray_src = cv2.cvtColor(src, cv2.COLOR_BGR2GRAY)
gray_src = cv2.bitwise_not(gray_src)
binary_src = cv2.adaptiveThreshold(gray_src, 255, cv2.ADAPTIVE_THRESH_MEAN_C, cv2.THRESH_BINARY, 15, -2)
# cv2.namedWindow("result image", cv2.WINDOW_AUTOSIZE)
# cv2.imshow("result image", binary_src)
# cv2.waitKey(0)
# 提取水平线
hline = cv2.getStructuringElement(cv2.MORPH_RECT, ((int(src.shape[1]*0.85)), 1), (-1, -1))
temp = cv2.erode(binary_src, hline)
dst = cv2.dilate(temp, hline)
dst = cv2.bitwise_not(dst)
# cv2.imshow("Final image", dst)
# cv2.waitKey(0)
lines = get_line_position(dst,backgroud_value=255)
# print(lines)
clear_line_img = clear_line(src,lines)
return clear_line_img
# clear_line_auto()
def cutting_img(src_img):
char_img_lists = []
h,w = src_img.shape[:2]
for i in range(4):
ROI_img = src_img[:, int(w * i / 4):int(w * (i+1) / 4)]
char_img_lists.append(ROI_img)
return char_img_lists
def clear_noise(image):
kernel = cv2.getStructuringElement(cv2.MORPH_RECT, (2, 2)) # 定义结构元素
closing = cv2.morphologyEx(image, cv2.MORPH_CLOSE, kernel) # 开运算
# cv2.imshow('morph_close',closing)
# cv2.waitKey(0)
# filename = join(ProcedurePath,img_name.split('.')[0]+ 'morph_close.png')
# cv2.imwrite(filename, closing)
# kernel = cv2.getStructuringElement(cv2.MORPH_RECT, (1, 1)) # 定义结构元素
# dilation = cv2.dilate(closing, kernel) # 腐蚀
# cv2.imshow('dilation',dilation)
# cv2.waitKey(0)
# filename = join(ProcedurePath, img_name.split('.')[0]+'morph_dilation.png')
# cv2.imwrite(filename, dilation)
# opening = cv2.morphologyEx(closing, cv2.MORPH_OPEN, kernel) # 开运算
# cv2.imshow('morph_open',opening)
# cv2.waitKey(0)
# filename = join(ProcedurePath, img_name.split('.')[0]+'morph_open.png')
# cv2.imwrite(filename, opening)
return closing
# return opening
def clear_backcolor_auto(img,frontcolor_num = 4):
Z = img.reshape((-1,3))
# convert to np.float32
Z = np.float32(Z)
j =0
# define criteria, number of clusters(K) and apply kmeans()
criteria = (cv2.TERM_CRITERIA_EPS + cv2.TERM_CRITERIA_MAX_ITER, 10, 1.0)
k = frontcolor_num
ret,label,center=cv2.kmeans(Z,k,None,criteria,50,cv2.KMEANS_RANDOM_CENTERS)
# Now convert back into uint8, and make original image
center = np.uint8(center)
couts=[]
for i in range(k):
backcolor = [0,0,0]
center = np.row_stack((center,backcolor))
# position = np.where(label==i,i,k)
couts.append(list(label.flatten()).count(i))
# res = center[position.flatten()]
# res2 = res.reshape((img.shape))
# # temp_img = position.reshape(img.shape[:2])
# # cv2.imshow("temp_img"+str(i),temp_img)
# cv2.imwrite(str(i+1)+'.png',res2)
# cv2.imshow("temp",res2)
# cv2.waitKey(0)
#print(couts)
backcolor_index = couts.index(max(couts))
#center = np.array([[0,0,0],[255,255,255]])
position = np.where(label==backcolor_index,np.uint8(255),np.uint8(0))
res = position.flatten()
res2 = res.reshape((img.shape[:2]))
# filename = join(ProcedurePath, img_name.split('.')[0] + '_clear_backcolor.png')
# cv2.imwrite(filename,res2)
# cv2.imshow('quondam image',img)
# cv2.waitKey(0)
# cv2.destroyAllWindows()
return res2
def divide_pic(Captcha_Path = TrainSet_Path,Divide_Path = Train_Divide_Path ,captcha_length=4,counts_file = "counts_file.txt"):
# 在路径下新建文件夹，名字为a-z,0-9,用于存储分割后的验证码
counts_recoder = {'total':0}
for i in range(26):
counts_recoder[chr(i+97)] = 0
if not os.path.exists(Divide_Path + chr(i+97)):
os.mkdir(Divide_Path + chr(i+97))
for j in range(10):
counts_recoder[str(j)] = 0
if not os.path.exists(Divide_Path + str(j)):
os.mkdir(Divide_Path + str(j))
fp = open(Captcha_Path + "/result.txt") # 验证码训练集的答案路径
divide_name = []
for x in fp.readlines():
divide_name.append(str(x).strip())
fp.close()
for i in range(1,len(os.listdir(Captcha_Path))):
src_img_name = str(i).zfill(4) + suffix
src_image = cv2.imread(join(Captcha_Path,src_img_name)) # 读取处理后的验证码
for j in range(captcha_length):# 每张验证码有四个字符
child_image = src_image[:,int(j*IMAGE_WEIGHT/4):int((j+1)*IMAGE_WEIGHT/4)] # 分割验证码图片（均分）
save_path = join(Divide_Path,divide_name[i-1][j])
save_img_name = str(i).zfill(4) + "-" + str(j) + suffix
cv2.imwrite(join(save_path,save_img_name),child_image) # 存储分割后的图片
counts_recoder[divide_name[i-1][j]]+=1
counts_recoder['total'] +=1
with open(join(Divide_Path,counts_file),'w+') as f:
for k,v in counts_recoder.items():
f.write(k+':'+str(v)+'\n')
def temp_test():
for _,_,files in os.walk(DataSet_Path):
for filename in files:
if filename.split('.')[1]!='png':
continue
# #获取降噪前图片
# clearbackcolor_img_path = join(ClearBackColor_Path,filename)
# clearbackcolor_img = cv2.imread(clearbackcolor_img_path)
# #测试函数处理后保存
# buffer_img_path =join(Buffer_Path,filename)
# buffer_img = clear_noise(clearbackcolor_img)
# cv2.imwrite(buffer_img_path,buffer_img)
#转为二值化图像（PIL型）
# need_img_path = join(Captcha_Path,filename)
# need_img = Image.open(need_img_path)
# need_img = need_img.convert('1')
# need_img.save(need_img_path)
if __name__ == "__main__":
temp_test()
divide_pic(Captcha_Path=TrainSet_Path,Divide_Path=Train_Divide_Path)
# divide_pic(Captcha_Path=TestSet_Path,Divide_Path=Test_Divide_Path)
input("pause")
for _,_,files in os.walk(DataSet_Path):
for filename in files:
if filename.split('.')[1]!='png':
continue
#获取原图像
img_path = join(DataSet_Path,filename)
img = cv2.imread(img_path)
#祛除干扰线
clearline_img_path = join(ClearLine_Path,filename)
clearline_img = clear_line_auto(img)
cv2.imwrite(clearline_img_path,clearline_img)
#祛除背景色
clearbackcolor_img_path = join(ClearBackColor_Path,filename)
clearbackcolor_img = clear_backcolor_auto(clearline_img)
cv2.imwrite(clearbackcolor_img_path,clearbackcolor_img)
#降噪处理
clearnoise_img_path = join(ClearNoise_Path,filename)
clearnoise_img = clear_noise(clearbackcolor_img)
cv2.imwrite(clearnoise_img_path,clearnoise_img)
#转为二值化图像（PIL型）
need_img_path = join(Captcha_Path,filename)
need_img = Image.Open(need_img_path)
need_img = need_img.convert('1')
need_img.save()