-
Notifications
You must be signed in to change notification settings - Fork 0
/
License_plate_detection_and_cut.py
244 lines (212 loc) · 8.66 KB
/
License_plate_detection_and_cut.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
175
176
177
178
179
180
181
182
183
184
185
186
187
188
189
190
191
192
193
194
195
196
197
198
199
200
201
202
203
204
205
206
207
208
209
210
211
212
213
214
215
216
217
218
219
220
221
222
223
224
225
226
227
228
229
230
231
232
233
234
235
236
237
238
239
240
241
242
243
244
import cv2 as cv
import numpy as np
import os
img_path = "E:/机器学习/较复杂环境下车牌号的识别/LPR/test_picture/test2.jpg"
save_path = "E:/机器学习/较复杂环境下车牌号的识别/test/test_save"
def show(name, img):
#显示图片
cv.namedWindow(str(name), cv.WINDOW_AUTOSIZE)
cv.imshow(str(name), img)
def binary(img):
#二值化处理去燥
for i in range(len(img)):
for j in range(len(img[0])):
if img[i][j]< 130:
img[i][j]=0
else:
img[i][j] = 255
return img
def separate_color_blue(img): # HSV阈值难以确定，暂时不用
#颜色提取
hsv = cv.cvtColor(img, cv.COLOR_BGR2HSV) # 色彩空间转换为hsv，便于分离
lower_hsv = np.array([105, 110, 115]) # 提取颜色的低值
high_hsv = np.array([130, 255,255]) # 提取颜色的高值
mask = cv.inRange(hsv, lowerb=lower_hsv, upperb=high_hsv)
mask = binary(mask)
# print("颜色提取完成")
return mask
def contour(img1,img2):
#检测轮廓
# img1 = cv.cvtColor(img1, cv.COLOR_BGR2GRAY)
# show("kkk",img1)
# cv.waitKey(0)
ret, img1 = cv.threshold(img1, 127, 255, cv.THRESH_BINARY)
image, contours, hier = cv.findContours(img1, cv.RETR_LIST, cv.CHAIN_APPROX_SIMPLE)
for c in contours: # 遍历轮廓
rect = cv.minAreaRect(c) # 生成最小外接矩形
h = min([int(rect[1][0]), int(rect[1][1])])
w = max([int(rect[1][0]), int(rect[1][1])])
box = cv.boxPoints(rect) # 计算最小面积矩形的坐标
box = np.int0(box) # 将坐标规范化为整数
cv.drawContours(img1, [box], 0, (255, 255, 255), 8)
# 补充完矩形框后，再检测一次，寻找车牌目标矩形
image, contours, hier = cv.findContours(img1, cv.RETR_LIST, cv.CHAIN_APPROX_SIMPLE)
count = 0
max_w = 0
max_h = 0
max_angle = 0
for c in contours: # 遍历轮廓
approx = cv.approxPolyDP(c, epsilon=5,closed=True)
rect = cv.minAreaRect(approx) # 生成最小外接矩形
h = min([int(rect[1][0]), int(rect[1][1])])
w = max([int(rect[1][0]), int(rect[1][1])])
box = cv.boxPoints(rect) # 计算最小面积矩形的坐标
box = np.int0(box) # 将坐标规范化为整数
cv.drawContours(img2, [box], 0, (0, 0, 255), 2)
# 只保留需要的轮廓
if (h > 1000 or w > 1000):
continue
if (h < 20 or w < 10):
continue
if (w / h > 5 or w / h <= 1.5):
continue
count += 1
angle = rect[2] # 获取矩形相对于水平面的角度
if angle > 0:
if abs(angle) > 45:
angle = 90 - abs(angle)
else:
if abs(angle) > 45:
angle = (90 - abs(angle))
# 筛选出最大的矩形对应的坐标
if w > max_w and h > max_h:
max_w, max_h = w, h
max_box = box.copy()
max_angle = angle
# print("angle", angle)
# print("坐标", box)
# show("img_1",img1)
# # cv.drawContours(img2, contours, -1, color=(0,0,255),thickness=2)
# show("img1_contours",img2)
# print("轮廓数量", count)
# cv.waitKey()
# print("max_box",max_box)
# if flag == False:
# cv.waitKey()
# show("max_box",img2)
if count==0:
print("Error, can not find License Plate!")
exit()
cv.drawContours(img2, [max_box], 0, (0, 255, 0), 10)
return img1,img2, max_box, max_angle
def rotate(img, angle):
#旋转图片
(h, w) = img.shape[:2] #获得图片高，宽
center = (w // 2, h // 2) #获得图片中心点
img_ratete = cv.getRotationMatrix2D(center, angle, 1)
rotated = cv.warpAffine(img, img_ratete, (w, h))
return rotated
# 裁剪车牌
def cut1(img, box,flag):
#从轮廓出裁剪图片
x, y=[], []
for i in range(len(box)):
x.append(box[i][0])
y.append(box[i][1])
x1, y1 = min(x), min(y) #获取左上角坐标
x2, y2 = max(x), max(y) #获取右下角坐标
x1, y1 = max([0,x1]), max([0,y1])
x2, y2 = max([0, x2]), max([0, y2])
# p为校验值
p = 0
if flag == False:
p = int(len(img) * 0.05)
img_cut = img[y1:y2, x1 + p:x2 - 2*p,:] #切片裁剪图像
return img_cut
# 裁剪出字符
def cut2(img_cut):
img_cut = cv.resize(img_cut, (440, 140))
img_cut = binary(cv.cvtColor(img_cut, cv.COLOR_BGR2GRAY))
img1 = img_cut[15:125, 15:61]
img2 = img_cut[15:125, 72:118]
img3 = img_cut[15:125, 151:197]
img4 = img_cut[15:125, 208:254]
img5 = img_cut[15:125, 265:311]
img6 = img_cut[15:125, 322:368]
img7 = img_cut[15:125, 379:425]
return img1,img2,img3,img4,img5,img6,img7
def cut_test_save():
bool = os.path.exists(save_path)
if bool == False:
os.makedirs(save_path)
# 解决imread不能读取中文路径
img = cv.imdecode(np.fromfile(img_path, dtype=np.uint8), flags=cv.IMREAD_COLOR)
img = cv.resize(img,(512,512))
img_separate = separate_color_blue(img.copy()) # 提取蓝色框先
cv.imencode('.png', img_separate)[1].tofile(save_path + '/' + 'test.jpg')
show("img_separate", img_separate)
img_contours2, img2, box, angle = contour(img_separate, img.copy()) # 轮廓检测，获取最外层矩形框的偏转角度
show("img2", img2)
show("img_contours2", img_contours2)
# img = cv.cvtColor(img,cv.COLOR_BGR2GRAY)
# img = binary(img)
img_cut = cut1(img.copy(), box,True)
show("img_cut",img_cut)
img_cut_rotate = rotate(img_cut, angle)
show("img_cut_rotate",img_cut_rotate)
img_cut_ = cv.resize(img_cut_rotate, (440, 140))
img_cut1, img_cut2, img_cut3, img_cut4, img_cut5, img_cut6, img_cut7 = cut2(img_cut_)
cv.imencode('.png', img_cut_rotate)[1].tofile(save_path + '/' + 'img_cut_rotate.png')
cv.imencode('.png', img_cut_)[1].tofile(save_path + '/' + 'img_cut.png')
cv.imencode('.png', img_cut1)[1].tofile(save_path + '/' + 'img_cut1.png')
cv.imencode('.png', img_cut2)[1].tofile(save_path + '/' + 'img_cut2.png')
cv.imencode('.png', img_cut3)[1].tofile(save_path + '/' + 'img_cut3.png')
cv.imencode('.png', img_cut4)[1].tofile(save_path + '/' + 'img_cut4.png')
cv.imencode('.png', img_cut5)[1].tofile(save_path + '/' + 'img_cut5.png')
cv.imencode('.png', img_cut6)[1].tofile(save_path + '/' + 'img_cut6.png')
cv.imencode('.png', img_cut7)[1].tofile(save_path + '/' + 'img_cut7.png')
show("img_cut1", img_cut1)
show("img_cut2", img_cut2)
show("img_cut3", img_cut3)
show("img_cut4", img_cut4)
show("img_cut5", img_cut5)
show("img_cut6", img_cut6)
show("img_cut7", img_cut7)
cv.waitKey(0)
cv.destroyAllWindows()
def detction_and_cut(img):
img = cv.resize(img, (512, 512))
img_separate = separate_color_blue(img.copy()) # 提取蓝色框先
try:
img_contours, img2, box, angle = contour(img_separate, img.copy()) # 轮廓检测，获取最外层矩形框的偏转角度
except ValueError:
print("未检测到车牌！")
return
img_cut = cut1(img.copy(), box,True)
img_cut_rotate = rotate(img_cut, angle)
img_cut1, img_cut2, img_cut3, img_cut4, img_cut5, img_cut6, img_cut7 = cut2(img_cut_rotate)
return [img_cut_rotate, img_cut1, img_cut2, img_cut3, img_cut4, img_cut5, img_cut6, img_cut7]
# 生成RGB色域，共16张图，每张图片大小为1024*1024
def RGB():
save_RGB_path = "./RGB3"
if os.path.exists(save_RGB_path)==False:
os.makedirs(save_RGB_path)
b,g,r = 0,0,0
for i in range(16):
img=[]
for j in range(1024):
w=[]
for k in range(1024):
h=[b,g,r]
if g < 255:
g += 1
elif r < 255:
r += 1
g = 0
elif b < 255:
b += 1
r = 0
g = 0
w.append(h)
img.append(w)
img = np.asarray(img)
cv.imencode('.jpg', img)[1].tofile(save_RGB_path + '/' + 'RGB_'+ str(i+1) +'.jpg')
if __name__ == "__main__":
cut_test_save()
# RGB()
# save_RGB_path = "./RGB"
# for file in os.listdir(save_RGB_path):
# filed = save_RGB_path + '/' + file
# img = cv.imdecode(np.fromfile(filed, dtype=np.uint8), flags=cv.IMREAD_COLOR)
# img2 = separate_color_blue(img)
# cv.imencode('.jpg', img2)[1].tofile(filed + '_blue' +'.jpg')