-
Notifications
You must be signed in to change notification settings - Fork 4
/
Copy pathprocess_captcha.py
131 lines (113 loc) · 3.82 KB
/
process_captcha.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
import cv2
import numpy as np
import os
import shutil
from matplotlib import pyplot as plt
path = os.path.dirname(os.path.realpath(__file__))
# print(os.path.join(path, 'pics'))
pics_src = os.path.join(path, 'pics')
parted_src = os.path.join(path, 'parted')
if os.path.isdir(parted_src):
shutil.rmtree(parted_src)
os.makedirs(parted_src)
num, dataset, x_train, y_train, x_test, y_test = 0, [], [], [], [], []
# letters = []
# falses = []
# from keras.models import model_from_json
#
# # load json and create model
# json_file = open('model.json', 'r')
# loaded_model_json = json_file.read()
# json_file.close()
# model = model_from_json(loaded_model_json)
# # load weights into new model
# model.load_weights("model.h5")
# print("Loaded model from disk")
pics = os.listdir(pics_src)
# print(pics)
for img_src in pics:
name = img_src.split('.')[0]
# if not os.path.isdir(parted_src + name):
# os.makedirs(parted_src + name)
img = cv2.imread(pics_src + img_src)
img = img[5:35, 5:145]
n_img = cv2.cvtColor(img, cv2.COLOR_BGR2GRAY)
color = ('b', 'g', 'r')
selected_color_max, max_hist = 'n', cv2.calcHist([n_img], [0], None, [256], [0, 256]).max()
# max_hist = 0
min_hist = 10000
selected_color_min = selected_color_max
# print(str('n') + ' ' + str(max_hist) + ' ' + name)
for i, col in enumerate(color):
histr = cv2.calcHist([img], [i], None, [256], [0, 256])
if histr.max() > max_hist:
max_hist = histr.max()
selected_color_max = col
if histr.max() < min_hist:
min_hist = histr.max()
selected_color_min = col
if max_hist < 1000:
selected_color = selected_color_min
else:
selected_color = selected_color_max
if selected_color == 'r':
img[:, :, 1] = 0
img[:, :, 2] = 0
elif selected_color == 'g':
img[:, :, 0] = 0
img[:, :, 2] = 0
elif selected_color == 'b':
img[:, :, 1] = 0
img[:, :, 0] = 0
img = cv2.cvtColor(img, cv2.COLOR_BGR2GRAY)
hist = cv2.calcHist([img], [0], None, [256], [0, 256])
print(name)
print(hist.max())
# plt.plot(hist)
# plt.xlim([0, 256])
# plt.show()
if hist.max() > 1100:
img = cv2.threshold(img, 0, 255, cv2.THRESH_BINARY_INV | cv2.THRESH_OTSU)[1]
else:
img = cv2.threshold(img, 0, 255, cv2.THRESH_BINARY | cv2.THRESH_OTSU)[1]
# for i in range(0, 5):
# if len(name) > 0:
# cv2.imwrite(parted_src + name + '/' + name[i] + '.png', img[:, max(i * 28 - 2, 1):min((i + 1) * 28 + 2, 139)])
# x_dataset.append(img)
# y_dataset.append(name[0:5])
if img.shape == (30, 140) and len(name) >= 5:
for j in range(0, 5):
img1 = np.zeros((30, 28, 1))
img1[:, :, 0] = img[:, j * 28: (j + 1) * 28]
if num > len(pics):
x_train.append(img1)
y_train.append(ord(name[0:5][j]) - 97)
else:
x_test.append(img1)
y_test.append(ord(name[0:5][j]) - 97)
# letters.append(name[0:5][j])
# if (name[0:5][j] == '￩'):
# falses.append(name)
num += 1
cv2.imwrite(parted_src + name + '.png', img)
# print(len(set(letters)))
# print((set(letters)))
# print(falses)
print(num)
x_train = np.array(x_train)
x_test = np.array(x_test)
y_train = np.array(y_train)
y_test = np.array(y_test)
dataset = (x_train, y_train), (x_test, y_test)
# prediction = list(model.predict(np.array([x_test[1], x_test[2]])))
# for predict in prediction:
# print(chr(list(predict).index(float(1))+97))
#
# cv2.imshow('image', x_test[1])
# cv2.waitKey(0)
# cv2.destroyAllWindows()
# cv2.imshow('image', x_test[2])
# cv2.waitKey(0)
# cv2.destroyAllWindows()
np.array(dataset).dump(open('captcha.npy', 'wb'))
# np.save('captcha', np.array(dataset))