CAPTCHACrack/Type1.py at master · programokey/CAPTCHACrack · GitHub

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
import pandas as pd
import skimage.io as io
import glob
from PIL import Image
import cv2
import shutil
import numpy as np
class Data(object):
    def __init__(self):
        self.label = {}
        df = pd.read_csv('sample_data/sample_submit_type_1.csv')
        for i, row in df.iterrows():
            self.label[row.IMG_ID] = row.GUESS
        print(self.label['0a0c7b9c-4e7b-11ea-a0ad-001a7dda7113'])
        print(len(self.label))
        dir = 'sample_data\\Type_1'
        for file in glob.glob(f'{dir}\\*.jpg'):
            id = file.split('\\')[-1].split('.')[0]
            if id not in self.label:
                print(id)
                shutil.copy(file, f'sample_data\\test\\{id}.jpg')
            #
        # self.load_imgs('sample_data/Type_1')

    def load_imgs(self, dir):
        for file in glob.glob(f'{dir}/*.jpg'):
            img = Image.open(file)
            print(img)


if __name__ == '__main__':
    # Data()
    # %%
    import cv2
    import numpy as np
    import matplotlib.pyplot as plt

    img = cv2.imread('sample_data/Type_1/0af7141e-4e7b-11ea-8c8b-001a7dda7113.jpg',cv2.IMREAD_COLOR)
    # img = cv2.imread('sample_data/Type_1/0af7141e-4e7b-11ea-8c8b-001a7dda7113.jpg', cv2.IMREAD_GRAYSCALE)
    # img = cv2.imread('sample_data/Type_1/0b1c83e6-4e7b-11ea-a116-001a7dda7113.jpg', 0)
    laplacian = cv2.Laplacian(img, -1)
    laplacian = cv2.convertScaleAbs(laplacian)/255
    laplacian = np.mean(np.square(laplacian), axis=-1)
    laplacian = laplacian/np.max(laplacian)
    plt.imshow(laplacian)
    plt.show()
    exit()
    # plt.imshow(img)
    # plt.show()

    # erosion = img
    # k = 3
    # kernel = np.ones((k, k), np.uint8)
    # erosion = cv2.erode(img, kernel, iterations=1)
    # plt.imshow(np.round(erosion).astype(np.uint8))
    # plt.show()

    blur = cv2.bilateralFilter(img, 7, 89, 50)
    thres, binary = cv2.threshold(blur, 82, 255, cv2.THRESH_BINARY_INV)

    # plt.imshow(binary)
    # res = cv2.blur(binary, (3, 3))
    # thres, binary = cv2.threshold(res, 200, 255, cv2.THRESH_BINARY)
    # binary = cv2.adaptiveThreshold(binary, 255, cv2.ADAPTIVE_THRESH_GAUSSIAN_C, \
    #                       cv2.THRESH_BINARY_INV, 11, 2)
    plt.imshow(binary)
    plt.show()
    exit()

    blur = cv2.bilateralFilter(img, 9, 75, 75)
    plt.imshow(blur)
    plt.show()

    thres, binary = cv2.threshold(blur, 82, 255, cv2.THRESH_BINARY_INV)
    # binary = cv2.adaptiveThreshold(img, 255, cv2.ADAPTIVE_THRESH_GAUSSIAN_C, \
    #                       cv2.THRESH_BINARY_INV, 11, 2)
    plt.imshow(cv2.blur(img, ksize=5))
    plt.show()
    exit()

    img = binary
    # k = 2
    # dilate_kernel = np.ones((k, k), np.uint8)/k**2
    # openning = cv2.dilate(img, dilate_kernel, iterations=1)
    # plt.imshow(openning)
    # plt.show()


    # def doCanny(x):
    #     # gauss = cv2.GaussianBlur(img, (3, 3), 1)
    #     position = cv2.getTrackbarPos("CannyBar", "Canny")
    #     # canny = cv2.Canny(binary, position, position * 5)
    #     canny = cv2.Laplacian(binary, ddepth=-1)
    #     cv2.imshow("Canny", canny)
    # cv2.namedWindow("Canny")
    # cv2.createTrackbar("CannyBar", "Canny", 1, 255, doCanny)
    # cv2.waitKey(0)
    # exit()


    # def do_thres(x):
    #     # gauss = cv2.GaussianBlur(img, (3, 3), 1)
    #     position = cv2.getTrackbarPos("ThresholdBar", "Canny")*255/100
    #     thres, binary = cv2.threshold(blur, position, maxval=255, type=cv2.THRESH_BINARY_INV)
    #     # canny = cv2.Canny(blur, position, position * 1.5)
    #     cv2.imshow("Canny", binary)
    # cv2.namedWindow("Canny")
    # cv2.createTrackbar("ThresholdBar", "Canny", 1, 100, do_thres)
    # cv2.waitKey(0)

    # canny = cv2.Canny(blur, 32*255/100, 32*255/100 * 1.5)
    # plt.imshow(canny)
    # plt.show()