-
Notifications
You must be signed in to change notification settings - Fork 0
/
interaction.py
168 lines (154 loc) · 5.96 KB
/
interaction.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
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
import socket
import sys
import os
import Image
import numpy as np
import glob
from scipy import signal, misc
from scipy.ndimage.morphology import binary_closing, binary_opening, binary_dilation, binary_erosion
from scipy.ndimage import sobel
from skimage.measure import label
from copy import deepcopy as dc
try:
from skimage import filters
except ImportError:
from skimage import filter as filters
IP_ADDRESS = '52.49.91.111'
PORT = 3456
BYTE_REQ = 10000000
PIXEL_REVERT = 40
'''
image_database = {}
for image in glob.glob('image_database/*'):
amount = int(os.path.split(image)[-1].split('.')[0])
image_content = misc.imread(image)
image_database[amount] = image_content
'''
image_database = {}
image_database[1] = (53, 53)
image_database[2] = (60, 60)
image_database[10] = (64, 64)
image_database[5] = (68, 68)
image_database[20] = (70, 70)
image_database[100] = (73, 73)
image_database[50] = (77, 77)
image_database[200] = (80, 80)
'''
def get_money_amount(image_matrix):
ret = 0
row, col, _ = image_matrix.shape
for (amount, pattern) in image_database.items():
prow, pcol, _ = pattern.shape
print row, col, prow, pcol
for i in range(row - prow + 1):
for j in range(col - pcol + 1):
delta = 0.0
for k in range(prow):
for l in range(pcol):
delta += np.linalg.norm(image_matrix[i, j, :]/255.0 - pattern[k, l, :]/255.0)
if delta < float(prow*pcol):
ret += 1
print ret
return ret
'''
def qmain():
image_content = misc.imread('imatge_3.jpeg')
print get_money_amount(image_content)
def main():
s = socket.socket(socket.AF_INET, socket.SOCK_STREAM)
s.settimeout(100)
s.connect((IP_ADDRESS, PORT))
current_level = 0
while True:
image_path = 'imatge_%d.jpeg'%current_level
f = open(image_path, 'wb')
current_level += 1
last_message = s.recv(BYTE_REQ)
while not 'money' in last_message:
print len(last_message)
f.write(last_message)
last_message = s.recv(BYTE_REQ)
# append the question, it wont hurt
f.write(last_message)
f.close()
preimg = misc.imread(image_path)
img = preimg[:, :, 0]
#img = signal.medfilt(img)
image_content = np.array(img, dtype = np.float) / 255.0
row, col = image_content.shape
# revert the image
for i in range(row):
for j in range(40, col, 80):
l = j
r = min(col-1, l+39)
env = r-l+1
for k in range(env//2):
tmp = image_content[i, l+k]
image_content[i, l+k] = image_content[i, l+env-k-1]
image_content[i, l+env-k-1] = tmp
print 'Money?'
thresh = filters.threshold_otsu(image_content)
to_erase = image_content < thresh
to_preserve = image_content >= thresh
image_content[to_erase] = 1.0
image_content[to_preserve] = 0.0
#image_content = binary_closing(image_content)
image_content = np.array(image_content, dtype=np.uint8)
labels = label(image_content)
label_count = np.max(labels)
answer = 0
# (h, w, id)
regions = []
total_coins = 0
for i in range(1, label_count+1):
xwhere, ywhere = np.where(labels == i)
height = max(xwhere) - min(xwhere) + 1
width = max(ywhere) - min(ywhere) + 1
if height < 50 or width < 50: continue
if height > 90 or width > 90: continue
regions.append((height, width, i, xwhere, ywhere))
regions = sorted(regions, key=lambda x: -x[0]*x[1])
for (height, width, regionid, xwhere, ywhere) in regions:
if np.sum(labels[min(xwhere):max(xwhere)+1, min(ywhere):max(ywhere)+1]) == 0:
print 'Discarded region'
continue
pixel_count = height * width
print 'Found a %d x %d region'%(height, width)
smallest_diff = 10**50
for (amount, (pwidth, pheight)) in image_database.items():
if amount == 50 or amount == 200:
half_x = (min(xwhere) + max(xwhere))//2
half_y = (min(ywhere) + max(ywhere))//2
quarter_amount = (width*height) / 4.0
white_ratio = np.sum(image_content[half_x:max(xwhere)+1, half_y:max(ywhere)+1]) / quarter_amount
print 'special coin %d found, white ratio is %f'%(amount, white_ratio)
thresh = 0.65
if amount == 50 and white_ratio > thresh:
continue
if (amount == 100 or amount == 200) and white_ratio < thresh:
continue
print 'coin is NOT rejected'
diff_x = abs(height - pheight) / float(height)
diff_y = abs(width - pwidth) / float(width)
diff = np.linalg.norm(np.array([diff_x, diff_y]))
#diff = abs((pixel_count - pheight*pwidth) / float(pheight*pwidth))
#diff = np.linalg.norm(np.array([height - pheight, width - pwidth]))
print 'Diff with %d coin: %f'%(amount, diff)
if diff < smallest_diff:
smallest_diff = diff
amount_to_add = amount
print '[IMPORTANT] Found a %d coin'%amount_to_add
answer += amount_to_add
total_coins += 1
labels[min(xwhere):max(xwhere)+1, min(ywhere):max(ywhere)+1] = 0
to_save = np.array(image_content*255.0, dtype = np.uint8)
misc.imsave(image_path.replace('jpeg', 'png'), to_save)
print str(answer)
s.send(str(answer)+'\n')
antwoord = s.recv(BYTE_REQ)
print 'Total coins: %d'%total_coins
print antwoord
if 'wrong' in antwoord:
exit(0)
if __name__ == '__main__':
main()