bank_ocr_inference.py
16.6 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
242
243
244
245
246
247
248
249
250
251
252
253
254
255
256
257
258
259
260
261
262
263
264
265
266
267
268
269
270
271
272
273
274
275
276
277
278
279
280
281
282
283
284
285
286
287
288
289
290
291
292
293
294
295
296
297
298
299
300
301
302
303
304
305
306
307
308
309
310
311
312
313
314
315
316
317
318
319
320
321
322
323
324
325
326
327
328
329
330
331
332
333
334
335
336
337
338
339
340
341
342
343
344
345
346
347
348
349
350
351
352
353
354
355
356
357
358
359
360
361
362
363
364
import base64
import os
import time
import cv2
import numpy as np
import requests
import tqdm
def image_to_base64(image):
image = cv2.imencode('.png', image)[1]
return image
def path_to_file(file_path):
f = open(file_path, 'rb')
return f
def bill_ocr(image):
f = image_to_base64(image)
resp = requests.post(url=r'http://139.196.149.46:9001/gen_ocr', files={'file': f})
results = resp.json()
ocr_results = results['ocr_results']
return ocr_results
def extract_minsheng_info(ocr_results):
name_prefix = '客户姓名:'
account_prefix = '客户账号:'
results = []
for value in ocr_results.values():
if name_prefix in value[1]:
if name_prefix == value[1]:
tmp_value, max_dis = [], 999999
top_right_x = value[0][2]
top_right_y = value[0][3]
for tmp in ocr_results.values():
if tmp[1] != name_prefix:
if abs(tmp[0][1] - top_right_y) < abs(value[0][3] - value[0][5]) / 2 and abs(
tmp[0][0] - top_right_x) < max_dis:
tmp_value = tmp
max_dis = abs(tmp[0][0] - top_right_x)
else:
continue
new_position = [value[0][0], value[0][1], tmp_value[0][2], tmp_value[0][3], tmp_value[0][4],
tmp_value[0][5],
value[0][6], value[0][7]]
results.append([value[1] + tmp_value[1], new_position])
else:
results.append([value[1], value[0]])
if account_prefix in value[1]:
if account_prefix == value[1]:
tmp_value, max_dis = [], 999999
top_right_x = value[0][2]
top_right_y = value[0][3]
for tmp in ocr_results.values():
if tmp[1] != account_prefix:
if abs(tmp[0][1] - top_right_y) < abs(value[0][3] - value[0][5]) / 2 and abs(
tmp[0][0] - top_right_x) < max_dis:
tmp_value = tmp
max_dis = abs(tmp[0][0] - top_right_x)
else:
continue
new_position = [value[0][0], value[0][1], tmp_value[0][2], tmp_value[0][3], tmp_value[0][4],
tmp_value[0][5],
value[0][6], value[0][7]]
results.append([value[1] + tmp_value[1], new_position])
else:
results.append([value[1], value[0]])
return results
def extract_gongshang_info(ocr_results):
name_prefix = '户名:'
account_prefix = '卡号:'
results = []
for value in ocr_results.values():
if name_prefix in value[1]:
if name_prefix == value[1]:
tmp_value, max_dis = [], 999999
top_right_x = value[0][2]
top_right_y = value[0][3]
for tmp in ocr_results.values():
if tmp[1] != name_prefix:
if abs(tmp[0][1] - top_right_y) < abs(value[0][3] - value[0][5]) / 2 and abs(
tmp[0][0] - top_right_x) < max_dis:
tmp_value = tmp
max_dis = abs(tmp[0][0] - top_right_x)
else:
continue
new_position = [value[0][0], value[0][1], tmp_value[0][2], tmp_value[0][3], tmp_value[0][4],
tmp_value[0][5],
value[0][6], value[0][7]]
results.append([value[1] + tmp_value[1], new_position])
else:
results.append([value[1], value[0]])
if account_prefix in value[1]:
if account_prefix == value[1]:
tmp_value, max_dis = [], 999999
top_right_x = value[0][2]
top_right_y = value[0][3]
for tmp in ocr_results.values():
if tmp[1] != account_prefix:
if abs(tmp[0][1] - top_right_y) < abs(value[0][3] - value[0][5]) / 2 and abs(
tmp[0][0] - top_right_x) < max_dis:
tmp_value = tmp
max_dis = abs(tmp[0][0] - top_right_x)
else:
continue
new_position = [value[0][0], value[0][1], tmp_value[0][2], tmp_value[0][3], tmp_value[0][4],
tmp_value[0][5],
value[0][6], value[0][7]]
results.append([value[1] + tmp_value[1], new_position])
else:
results.append([value[1], value[0]])
return results
def extract_zhongguo_info(ocr_results):
name_prefix = '客户姓名:'
account_prefix = '借记卡号:'
results = []
for value in ocr_results.values():
if name_prefix in value[1]:
if name_prefix == value[1]:
tmp_value, max_dis = [], 999999
top_right_x = value[0][2]
top_right_y = value[0][3]
for tmp in ocr_results.values():
if tmp[1] != name_prefix:
if abs(tmp[0][1] - top_right_y) < abs(value[0][3] - value[0][5]) / 2 and abs(
tmp[0][0] - top_right_x) < max_dis:
tmp_value = tmp
max_dis = abs(tmp[0][0] - top_right_x)
else:
continue
new_position = [value[0][0], value[0][1], tmp_value[0][2], tmp_value[0][3], tmp_value[0][4],
tmp_value[0][5],
value[0][6], value[0][7]]
results.append([value[1] + tmp_value[1], new_position])
else:
results.append([value[1], value[0]])
if account_prefix in value[1]:
if account_prefix == value[1]:
tmp_value, max_dis = [], 999999
top_right_x = value[0][2]
top_right_y = value[0][3]
for tmp in ocr_results.values():
if tmp[1] != account_prefix:
if abs(tmp[0][1] - top_right_y) < abs(value[0][3] - value[0][5]) / 2 and abs(
tmp[0][0] - top_right_x) < max_dis:
tmp_value = tmp
max_dis = abs(tmp[0][0] - top_right_x)
else:
continue
new_position = [value[0][0], value[0][1], tmp_value[0][2], tmp_value[0][3], tmp_value[0][4],
tmp_value[0][5],
value[0][6], value[0][7]]
results.append([value[1] + tmp_value[1], new_position])
else:
results.append([value[1], value[0]])
return results
def extract_jianshe_info(ocr_results):
name_prefixes = ['客户名称:', '户名:']
account_prefixes = ['卡号/账号:', '卡号:']
results = []
for value in ocr_results.values():
for name_prefix in name_prefixes:
if name_prefix in value[1]:
if name_prefix == value[1]:
tmp_value, max_dis = [], 999999
top_right_x = value[0][2]
top_right_y = value[0][3]
for tmp in ocr_results.values():
if tmp[1] != name_prefix:
if abs(tmp[0][1] - top_right_y) < abs(value[0][3] - value[0][5]) / 2 and abs(
tmp[0][0] - top_right_x) < max_dis:
tmp_value = tmp
max_dis = abs(tmp[0][0] - top_right_x)
else:
continue
new_position = [value[0][0], value[0][1], tmp_value[0][2], tmp_value[0][3], tmp_value[0][4],
tmp_value[0][5],
value[0][6], value[0][7]]
results.append([value[1] + tmp_value[1], new_position])
break
else:
results.append([value[1], value[0]])
break
for account_prefix in account_prefixes:
if account_prefix in value[1]:
if account_prefix == value[1]:
tmp_value, max_dis = [], 999999
top_right_x = value[0][2]
top_right_y = value[0][3]
for tmp in ocr_results.values():
if tmp[1] != account_prefix:
if abs(tmp[0][1] - top_right_y) < abs(value[0][3] - value[0][5]) / 2 and abs(
tmp[0][0] - top_right_x) < max_dis:
tmp_value = tmp
max_dis = abs(tmp[0][0] - top_right_x)
else:
continue
new_position = [value[0][0], value[0][1], tmp_value[0][2], tmp_value[0][3], tmp_value[0][4],
tmp_value[0][5],
value[0][6], value[0][7]]
results.append([value[1] + tmp_value[1], new_position])
break
else:
results.append([value[1], value[0]])
break
return results
def extract_nongye_info(ocr_results):
name_prefixes = ['客户名:', '户名:']
account_prefixes = ['账号:']
results = []
is_account = True
for value in ocr_results.values():
for name_prefix in name_prefixes:
if name_prefix in value[1] and account_prefixes[0][:-1] not in value[1]:
if name_prefix == value[1]:
tmp_value, max_dis = [], 999999
top_right_x = value[0][2]
top_right_y = value[0][3]
for tmp in ocr_results.values():
if tmp[1] != name_prefix:
if abs(tmp[0][1] - top_right_y) < abs(value[0][3] - value[0][5]) / 2 and abs(
tmp[0][0] - top_right_x) < max_dis:
tmp_value = tmp
max_dis = abs(tmp[0][0] - top_right_x)
else:
continue
new_position = [value[0][0], value[0][1], tmp_value[0][2], tmp_value[0][3], tmp_value[0][4],
tmp_value[0][5],
value[0][6], value[0][7]]
results.append([value[1] + tmp_value[1], new_position])
break
else:
results.append([value[1], value[0]])
break
if name_prefix in value[1] and account_prefixes[0][:-1] in value[1] and len(value[1].split(":")[0]) <= 5:
is_account = False
if len(value[1]) == 5:
tmp_value, max_dis = [], 999999
top_right_x = value[0][2]
top_right_y = value[0][3]
tmp_info = {}
for tmp in ocr_results.values():
if tmp[1] != value[1]:
if abs(tmp[0][1] - top_right_y) < abs(value[0][3] - value[0][5]) / 2:
tmp_info[abs(tmp[0][0] - top_right_x)] = tmp
else:
continue
tmp_info_id = sorted(tmp_info.keys())
if not tmp_info[tmp_info_id[0]][1].isdigit() and len(tmp_info[tmp_info_id[0]][1]) > 19:
tmp_value = tmp_info[tmp_info_id[0]]
new_position = [value[0][0], value[0][1], tmp_value[0][2], tmp_value[0][3], tmp_value[0][4],
tmp_value[0][5],
value[0][6], value[0][7]]
results.append([value[1] + tmp_value[1], new_position])
if tmp_info[tmp_info_id[0]][1].isdigit():
tmp_value = tmp_info[tmp_info_id[1]]
new_position = [value[0][0], value[0][1], tmp_value[0][2], tmp_value[0][3], tmp_value[0][4],
tmp_value[0][5],
value[0][6], value[0][7]]
results.append([value[1] + tmp_value[1], new_position])
break
elif len(value[1]) < 25:
tmp_info = {}
top_right_x = value[0][2]
top_right_y = value[0][3]
for tmp in ocr_results.values():
if tmp[1] != value[1]:
if abs(tmp[0][1] - top_right_y) < abs(value[0][3] - value[0][5]) / 2:
tmp_info[abs(tmp[0][0] - top_right_x)] = tmp
else:
continue
tmp_info_id = sorted(tmp_info.keys())
tmp_value = tmp_info[tmp_info_id[0]]
new_position = [value[0][0], value[0][1], tmp_value[0][2], tmp_value[0][3], tmp_value[0][4],
tmp_value[0][5],
value[0][6], value[0][7]]
results.append([value[1] + tmp_value[1], new_position])
break
else:
results.append([value[1], value[0]])
break
if is_account:
for account_prefix in account_prefixes:
if account_prefix in value[1]:
if account_prefix == value[1]:
tmp_value, max_dis = [], 999999
top_right_x = value[0][2]
top_right_y = value[0][3]
for tmp in ocr_results.values():
if tmp[1] != account_prefix:
if abs(tmp[0][1] - top_right_y) < abs(value[0][3] - value[0][5]) / 2 and abs(
tmp[0][0] - top_right_x) < max_dis:
tmp_value = tmp
max_dis = abs(tmp[0][0] - top_right_x)
else:
continue
new_position = [value[0][0], value[0][1], tmp_value[0][2], tmp_value[0][3], tmp_value[0][4],
tmp_value[0][5],
value[0][6], value[0][7]]
results.append([value[1] + tmp_value[1], new_position])
break
else:
results.append([value[1], value[0]])
break
else:
break
return results
def extract_bank_info(ocr_results):
results = []
for value in ocr_results.values():
if value[1].__contains__('建设'):
results = extract_jianshe_info(ocr_results)
break
elif value[1].__contains__('民生'):
results = extract_minsheng_info(ocr_results)
break
elif value[1].__contains__('农业'):
results = extract_nongye_info(ocr_results)
break
elif value[1].__contains__('中国银行'):
results = extract_zhongguo_info(ocr_results)
break
if len(results) == 0:
results = extract_gongshang_info(ocr_results)
return results
if __name__ == '__main__':
path = '/data/situ_invoice_bill_data/new_data/qfs_bank_bill_data/minsheng/authentic/images/val'
save_path='/data/situ_invoice_bill_data/new_data/results'
bank='minsheng'
if not os.path.exists(os.path.join(save_path,bank)):
os.makedirs(os.path.join(save_path,bank))
save_path=os.path.join(save_path,bank)
for j in tqdm.tqdm(os.listdir(path)):
# if True:
img=cv2.imread(os.path.join(path,j))
# img = cv2.imread('/data/situ_invoice_bill_data/new_data/results/nongye/6/_1597382769.6449914page_23_img_0.jpg')
st = time.time()
ocr_result = bill_ocr(img)
et1 = time.time()
result = extract_bank_info(ocr_result)
et2 = time.time()
for i in range(len(result)):
cv2.rectangle(img, (result[i][1][0], result[i][1][1]), (result[i][1][4], result[i][1][5]), (0, 0, 255), 2)
# cv2.imshow('img',img)
# cv2.waitKey(0)
cv2.imwrite(os.path.join(save_path,j),img)
print('spend:{} ocr:{} extract:{}'.format(et2 - st, et1 - st, et2 - et1))