CamScanner/camscanner.py at master · shloknangia/CamScanner · 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
119
120
121
122
123
124
125
126
127
import cv2
import numpy as np
import matplotlib.pyplot as plt

image = cv2.imread('./sample-1.jpg')
print image.shape
plt.figure(0)
plt.imshow(image, cmap = 'gray')
plt.figure(12)
# plt.imshow(rotate_bound(image, 180))
# plt.imshow(image)
# plt.show()

image = cv2.resize(image, (1500, 880))
orig = image.copy()
print image.shape
# plt.imshow(image)
# plt.show()


def rectify(h):
	# print h
	h = h.reshape((4, 2))
	# print h
	hnew =  np.zeros((4, 2), dtype = np.float32)
	add = h.sum(1)
	# print "add:", add
	hnew[0] = h[np.argmin(add)]
	hnew[2] = h[np.argmax(add)]

	diff = np.diff(h, axis = 1)
	# print "diff:", diff
	hnew[1] = h[np.argmin(diff)]
	hnew[3] = h[np.argmax(diff)]

	return hnew


# edge detection

# convert to grayscale
gray = cv2.cvtColor(image, cv2.COLOR_BGR2GRAY)
plt.figure(1, figsize = (7, 7))
plt.imshow(gray, cmap = 'gray')

# blurring , gaussian blur
blurred = cv2.GaussianBlur(gray, (5, 5), 0)
# blurred = cv2.medianBlur(gray, 5)
plt.figure(2, figsize = (7, 7))
plt.imshow(blurred, cmap = 'gray')

# applying edge detection
edged = cv2.Canny(blurred, 0, 50)
plt.figure(3, figsize = (7, 7))
plt.imshow(edged, cmap = 'gray')

# plt.show()

# finding largest contour in edgeded image

# find contour
(_, contours, _) = cv2.findContours(edged, cv2.RETR_LIST, cv2.CHAIN_APPROX_NONE)

# sort the contours by ares in decending order
contours = sorted(contours, key = cv2.contourArea, reverse = True)

# plot boundary against largest one
x, y, w, h = cv2.boundingRect(contours[1])
cv2.rectangle(image, (x, y), (x+w, y+h), (0, 255, 0), 2)
plt.figure(4, figsize = (7, 7))
plt.imshow(image, cmap = 'gray')
# plt.show()

# largest contour with 4 vertices
for c in contours:
	p = cv2.arcLength(c, True)
	approx = cv2.approxPolyDP(c, 0.02 * p, True)

	if len(approx) == 4:
		target = approx
		break

print "largest contour:", target

# plotting it
cv2.drawContours(image, [target], -1, (255, 0, 0), 2)
plt.figure(5, figsize = (7, 7))
plt.imshow(image, cmap = 'gray')
# plt.show()

# mapping target points to a quadrilateral

print "rectify called:"
approx = rectify(target)
print approx

pts2 = np.float32([[0, 0], [800, 0], [800, 800], [0, 800]])
M = cv2.getPerspectiveTransform(approx, pts2)
dst = cv2.warpPerspective(orig, M, (800, 800))

dst = cv2.cvtColor(dst, cv2.COLOR_BGR2GRAY)
plt.figure(6, figsize = (7, 7))
plt.imshow(dst, cmap = 'gray')
# plt.show()

# using thresholding on wrapped image to get scanned effect

ret, th1 = cv2.threshold(dst, 127, 255, cv2.THRESH_BINARY)
plt.figure(7, figsize = (7, 7))
plt.imshow(th1, cmap = 'gray')

th2 = cv2.adaptiveThreshold(dst, 255, cv2.ADAPTIVE_THRESH_MEAN_C, cv2.THRESH_BINARY, 11, 2)
plt.figure(8, figsize = (7, 7))
plt.imshow(th2, cmap = 'gray')

th3 = cv2.adaptiveThreshold(dst, 255, cv2.ADAPTIVE_THRESH_GAUSSIAN_C, cv2.THRESH_BINARY, 11, 2)
plt.figure(9, figsize = (7, 7))
plt.imshow(th3, cmap = 'gray')

ret2,th4 = cv2.threshold(dst, 0, 255, cv2.THRESH_BINARY+cv2.THRESH_OTSU)
plt.figure(10, figsize = (7, 7))
plt.imshow(th4, cmap = 'gray')

plt.show()