-
Notifications
You must be signed in to change notification settings - Fork 0
/
Image.py
174 lines (152 loc) · 5.83 KB
/
Image.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
169
170
171
172
173
174
import cv2
import numpy as np
class Image:
HEIGHT = 550
WIDTH = 350
img = 0
remove_back_img = 0
number_area = 0
def __init__(self,img):
self.img = img
H,W,_ = img.shape
if H/W > 0.6 and H/W < 0.69:
self.remove_back_img = cv2.resize(self.img,(self.HEIGHT,self.WIDTH))
else:
self.remove_back_img = self.removeBackground()
self.number_area = self.position(self.remove_back_img)
self.pos_img = self.getNumberArea()
def getNumberArea(self):
num_img = self.number_area
h,w,_ = num_img.shape
gray_img = cv2.cvtColor(num_img,cv2.COLOR_BGR2GRAY)
dilate_img= self.embossment(gray_img)
embo_img = cv2.medianBlur(dilate_img,3)
_,thresh_img= cv2.threshold(embo_img,155,255,cv2.THRESH_BINARY)
thresh_img = cv2.medianBlur(thresh_img,3)
thresh_img = cv2.GaussianBlur(thresh_img,(3,3),0)
thresh_img = cv2.dilate(thresh_img,None,iterations=10)
_,thresh_img= cv2.threshold(thresh_img,220,255,cv2.THRESH_BINARY)
a = np.zeros(w,np.uint8)
for j in range(0,w): #计算水平方向上的黑色像素点数目,
for i in range(0,h):
if thresh_img[i,j] == 0:
a[j] += 1
a = a[::-1]
length = int(0.75 * w)
min_ = sum(a)
start = 0
for i in range(len(a)):
if a[i] < 15:
a[i] = 0
else:
a[i] = 35
for i in range(w - length):
end = i + length
mean_ = a[i:end].mean()
if(min_ > mean_ and i < 50):
min_ = mean_
start = i
end = w - start
a = a[::-1]
min_ = sum(a)
start = 0
for i in range(130):
mean_ = a[i:end].mean()
if min_ > mean_:
min_ = mean_
start = i
# print(start,end)
self.W_end = end + 20
self.W_start = start + 25
# cv2.imshow('aaa',num_img[:,start:end+10])
return num_img[:,start+5:end]
def removeBackground(self):
resize_img = cv2.resize(self.img,(self.HEIGHT,self.WIDTH),0,0,cv2.INTER_NEAREST) #调整图片大小
self.img = resize_img
gray_img = cv2.cvtColor(resize_img,cv2.COLOR_BGR2GRAY) #灰度处理
blur_img = cv2.medianBlur(gray_img,9) #中值滤波去除噪声
x = cv2.Sobel(blur_img,cv2.CV_32F,1,0,3) #Sobel边缘检测
y = cv2.Sobel(blur_img,cv2.CV_32F,0,1,3)
absX = cv2.convertScaleAbs(x)
absY = cv2.convertScaleAbs(y)
sobel_img = cv2.addWeighted(absX,0.5,absY,0.5,0)
thresh_img = cv2.adaptiveThreshold(sobel_img,255,cv2.ADAPTIVE_THRESH_GAUSSIAN_C,cv2.THRESH_BINARY,3,0) #自适应二值化
cnts,_ = cv2.findContours(thresh_img,cv2.RETR_EXTERNAL,cv2.CHAIN_APPROX_SIMPLE) #找到最大连通区域
temp = 0
W = 0
H = 0
X = 0
Y = 0
for i in range(0,len(cnts)):
x,y,w,h = cv2.boundingRect(cnts[i])
if(temp <w + h):
temp = w+h
W = w
H = h
X = x
Y = y
remove_back_img = resize_img[Y:Y+H,X:X+W]
return cv2.resize(remove_back_img,(self.HEIGHT,self.WIDTH),cv2.INTER_NEAREST)
def embossment(self,img):
H,W = img.shape
dst = np.zeros((H,W),np.uint8)
for i in range(0,H):
for j in range(0,W-1):
grayP0 = int(img[i,j])
grayP1 = int(img[i,j+1])
newP = grayP0 - grayP1 + 150
if newP > 255:
newP = 255
if newP < 0:
newP = 0
dst[i,j] = newP
return dst
def horizontal(self,img):
H,W = img.shape
# test_img = np.ones((H,W)) * 255
hor_array = np.zeros(H,np.int32)
for j in range(0,H):
for i in range(0,W):
if img[j,i]== 0 :
hor_array[j]+=1
# test_img[j,i] = 255
# for j in range(0,H):
# for i in range(0,hor_array[j]):
# test_img[j,i]=0
return hor_array
# return hor_array,test_img
def getArea(self,array):
H = len(array)
label_H = int(H / 10)
min_ = sum(array)
ans = 0
for i in range(int(1/2 * H) - label_H): #从图像高2/5位置处开始进行平均值计算。
a = int(2/5 * H) + i
b = int(2/5 * H) + i + label_H
mean = array[a:b].mean()
if mean < min_:
ans = a
min_ = mean
if a > 0.6 * H:
return ans,ans+label_H
def position(self,img):
gray_img = cv2.cvtColor(img, cv2.COLOR_BGR2GRAY)
gray_img = cv2.dilate(gray_img,None,iterations=2)
gray_img = cv2.erode(gray_img,None,iterations=2)
emboss_img = self.embossment(gray_img)
sobel_x = cv2.Sobel(emboss_img,cv2.CV_32F,1,0,3) #边缘检测
sobel_y = cv2.Sobel(emboss_img,cv2.CV_32F,0,1,3)
absX = cv2.convertScaleAbs(sobel_x)
absY = cv2.convertScaleAbs(sobel_y)
sobel_img = cv2.addWeighted(absX,0.5,absY,0.5,0)
sobel_img = cv2.medianBlur(sobel_img,11) #中值模糊
sobel_img = cv2.dilate(sobel_img,None,iterations=2) #膨胀
_,threshold = cv2.threshold(sobel_img,10,255,cv2.THRESH_BINARY) #二值化
threshold = cv2.GaussianBlur(threshold,(9,9),0) #高斯模糊
# pixel_array,test_img = self.horizontal(threshold)
pixel_array = self.horizontal(threshold) #对图形黑色像素进行竖直投影
start,end = self.getArea(pixel_array)
self.H_start = start
self.H_end = end
res = img[start:end,20:]
return res