-
Notifications
You must be signed in to change notification settings - Fork 0
/
main.py
145 lines (113 loc) · 5.31 KB
/
main.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
import cv2 as cv
import numpy as np
from timeit import default_timer as timer
import card_finder as cf
import math
# %% WCZYTANIE OBRAZU
print("Start")
start_p=timer()
plik='013_blur.jpg'#<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<WYBÓR PLIKU
img_full = cv.imread(plik,0)
# 013.jpg, 013_blur.jpg, 013_gradient.jpg 013_salt.jpg
img_full_color=cv.imread(plik,1)
scale= 24 ## zmniejszenie rozmiaru pikseli
newX,newY = img_full.shape[1]/scale, img_full.shape[0]/scale
img = cv.resize(img_full,(int(newX),int(newY)))
cf.show(img,"obraz wejsciowy")
# %% ZNAJDOWANIE KART
#img=operation(img,get_circle(2),'median') #USUWA SZUM, NIEPOTRZEBNE
thresh1=cf.thresh(img,130)### BINARYZACJA
cf.show(thresh1,"Po binaryzacji")
background=cf.select_object(thresh1,(0,1),None) #ZAZNACZENIE CAŁEGO TŁA
cards_mask=np.logical_not(background["mask_img"]).astype(np.uint8) #KARTY
cards_list=cf.filter_object(cards_mask,500) #OPERACJA SEGMENTACJI
for i,card in enumerate(cards_list): #TYLKO DO ZAZNACZENIA KART
ctr=tuple(map(int,cf.center(card["mask_img"])))
card["center"]=ctr
card["index"]=i
cv.circle(cards_mask,ctr,3,(0,0,0))
cv.putText(cards_mask,str(i),ctr,cv.FONT_HERSHEY_SIMPLEX,1,(0,0,0))
cf.show(cards_mask,'Wszystkie potencjalne karty wyczyszczone')
# %% OPERACJA DLA DANEJ KARTY
for card in cards_list:
cf.show(card["mask_img"],f'maska')
kernel=np.array([[1,1,1],
[1,-8,1],
[1,1,1]])
card_border1=cf.operation(card["mask_img"],kernel,"convolution") #WYKRYCIE KRAWEDZI
card_border1=cf.thresh(abs(card_border1))
#show(card_border1,f"Krawędź karty Laplace {nr}") #ALBO TAK:
#card_border=operation(karta2,get_circle(3),"dilate")-karta2
cf.show(card_border1,f"Krawędź karty")
lines=cf.line_transform(card_border1,(400,400)) ##TRANSFROMACJA
cf.show(lines,"Hough")
points=cf.get_lines(lines,4,30,card_border1.shape)
#FUNKJA ZNAJDUJE DANA ILOSC MAKSIMOW LOKALNYCH W OBRAZIE TRANSFORMATY
#points zawiera listę puntów reprezentującyh proste na obrazie
corners=cf.find_corners(points["ab"])
card["corners"]=corners
card["eges"]=cf.sort_lines(corners)
card["angle"]=math.degrees(math.atan(2/(card["eges"][0][0]+
card["eges"][1][0])))
img_draw=img.copy() #WIZUALIZACJA NAROZNIKOW NA KARCE
for i, p in enumerate(corners):
cv.putText(img_draw,str(i), (int(p[0]),int(p[1])),
cv.FONT_HERSHEY_SIMPLEX,1,(0,0,0))
cv.circle(img_draw, (int(p[0]),int(p[1])),4, (255,0,0))
for point_ar in points["ar"]: #WIZUALIZACJA PROSTYCH W TRANSFORMACIE
cv.circle(lines, tuple(map(int,point_ar)),4, (255,0,0))
for point_ab in points["ab"]: #WIZUALIZACJA PROSTYCH NA OBRAZIE
a,b=point_ab
h,w=img.shape
cv.line(img_draw,(0,int(b)),(int(w),int(a*w+b)),(255,0,0))
cf.show(lines,"Wykryte proste")
cf.show(img_draw,"Wykryte boki karty")
# input("Wcisnij dowolny klawisz...")
# %%
cards_list=sorted(cards_list,key=lambda i:abs(i['angle']),reverse=True)
print("Znalezione karty:")
for card in cards_list:
print(f"nr:{card['index']} w srodku {card['center']} o nachyleniu do OY {card['angle']:.2f} stopni","\n"
f"proste zawierające dłuższe boki:\n",
f"y={card['eges'][0][0]:.2f}x+{card['eges'][0][1]:.2f}","\n"
f"y={card['eges'][1][0]:.2f}x+{card['eges'][1][1]:.2f}")
# %% ANALIZA OBRAZU DANEJ KARTY
print(f"Rozpoznawanie karty nr {cards_list[0]['index']}")
width,height=130,200
warped=np.zeros((height,width))
rectangle=np.array([[0,0],[width,0],[width,height],[0,height]],dtype=np.float32)
source=np.array(cards_list[0]['corners'],dtype=np.float32)*scale #narozniki na glownym obrazie
matrix=cv.getPerspectiveTransform(source, rectangle)
warped=cv.warpPerspective(img_full, matrix, (width,height))
cf.show(warped,"karta bez perspetywy")
border=20
roi=warped[border:-border,border:-border]
median=np.median(warped)
roi_t=cf.thresh(roi,0.65*median) #binaryzacja jednej karty
roi_t=cf.operation(roi_t,cf.get_circle(2),"median")#flitr medianowy
cf.show(roi_t,"Tu szukam symboli")
#cv.imwrite("trefl.png",roi_t*255)
symbols=cf.filter_object(np.logical_not(roi_t).astype(np.uint8), 120) #SEGMEMTACJA SYMBOLI
for symbol in symbols:
symbol["center"]=cf.center(symbol["mask_img"])
patterns=["kier","pik","trefl","karo"]
corr=dict.fromkeys(patterns,0)
for pattern in patterns:
img=np.logical_not(cv.imread(pattern+".png",0))*2-1
card=np.logical_not(roi_t)
result=cf.operation(card,img,'convolution')
result_flip=cf.operation(card,np.flip(img),'convolution')#ZBADANIE KORELACJI
#cf.show(result, f"Porownanie z {pattern}")
corr[pattern]=max(result.max(),result_flip.max())
type_of_card=max(corr.keys(),key=lambda x:corr[x])
print(f"Ta karta to {len(symbols)} {type_of_card}")
# %%
matrix_inv=np.linalg.inv(matrix)
roi_t_drw=roi_t.copy()
img_full_color=cv.cvtColor(img_full,cv.COLOR_GRAY2RGB)
for symbol in symbols:
center=np.dot(matrix_inv,np.append((symbol['center']+np.array([border,border])),1))
cv.circle(img_full_color,tuple(map(int,center[:2]/center[2])),20,(255,0,0),thickness=15)
cv.circle(warped,tuple(map(int,symbol['center']+np.array([border,border]))),1,255,thickness=2)
cf.show(warped,"wynik")
cf.show(img_full_color,'Wynik')