-
Notifications
You must be signed in to change notification settings - Fork 0
/
Center.py
299 lines (249 loc) · 11 KB
/
Center.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
245
246
247
248
249
250
251
252
253
254
255
256
257
258
259
260
261
262
263
264
265
266
267
268
269
270
271
272
273
274
275
276
277
278
279
280
281
282
283
284
285
286
287
288
289
290
291
292
293
294
295
296
297
298
299
import cv2
import numpy as np
from math import *
# from CMDcontrol import action_append,action_list
from Robot_control import action_append,action_list
from Undistort import undistort_chest
from Video_stream import LoadStreams
import time
def Calculate_position(line):
[ub, ut] = line
if ub == ut:
ub = ub + 1
# (ut, ub) = (484,-360)
# input variables ##########################################################
# biaoding photo
(u, v, ang1, ang2, ang3, h) = (243,189,-1.5479,-0.5800,1.5526,20.6104)
# inside camera ############################################################
f = 200 # pixel
x_fen = 480
y_fen = 640
# find lanes jiaodian in photo #############################################
# rotation matrix of camera: R1 rotates first
R1 = np.asarray([[cos(ang1), sin(ang1), 0], [-sin(ang1), cos(ang1), 0], [0, 0, 1]], dtype=float)
R2 = np.asarray([[cos(ang2), 0, -sin(ang2)], [0, 1, 0], [sin(ang2), 0, cos(ang2)]], dtype=float)
R3 = np.asarray([[cos(ang3), sin(ang3), 0], [-sin(ang3), cos(ang3), 0], [0, 0, 1]], dtype=float)
# Zc[0 0 1]axis vector in no-rotate cam coordinate system: Zc[x y z]
Zc = np.matmul(np.matmul(np.matmul(R3,R2),R1),np.matrix([[0], [0], [1]]))
# find projection vector of Zc in Xc0-Zc0 plane
Zcproj = np.asarray([Zc[0][0], 0, Zc[2][0]], dtype=float)
# find jiaoxian vector of Xc0-Zc0 plane and photo
XZplane = np.asarray([-Zc[2][0], 0, Zc[0][0]], dtype=float)
# Xc[1 0 0]axis vector in no-rotate cam coordinate system: Xc[x y z]
Xc = np.matmul(np.matmul(np.matmul(R3,R2),R1),np.matrix([[1], [0], [0]]))
# calculate absolute jiajiao of Xc and XZplane
Xc.shape = (3,1)
XZplane.shape = (1,3)
# print(np.matmul(XZplane, Xc))
cosXZ = abs( np.matmul(XZplane, Xc)/np.linalg.norm(Xc, ord=2)/np.linalg.norm(XZplane, ord=2) )
aXZ = np.sign(-Xc[0][0]*Xc[1][0]) * acos(cosXZ) # aXZ+, XZplane youdi zuogao in photo
tanXZ = tan(aXZ)
# calculate jiaodian: (ud, vd)
# original expression:
# (vd-v)/(ud-u) = tanXZ
# (ut-ud)/vd = (ud-ub)/(y_fen-vd)
ud = ( ut*y_fen + (ut-ub)*(tanXZ*u - v) ) / ( tanXZ*(ut-ub) + y_fen)
vd = ( tanXZ*y_fen*(ut-u) + y_fen*v ) / ( tanXZ*(ut-ub) + y_fen)
# calculate y-rotation of the car ##########################################
# vector of camdevi(u,v)-camorigin
vecuv = np.asarray([x_fen/2, y_fen/2, f], dtype=float) - np.asarray([u, v, 0], dtype=float)
vecuvd = np.asarray([x_fen/2, y_fen/2, f], dtype=float) - np.asarray([ud, vd, 0], dtype=float)
vecuv.shape = (1,3)
vecuvd.shape = (3,1)
# calculate absolute jiajiao of two vectors
cosyd = np.matmul(vecuv,vecuvd)/np.linalg.norm(vecuv, ord=2)/np.linalg.norm(vecuvd, ord=2)
ayd = np.sign(u-ud) * acos(cosyd) # ud < u, ayd+, car rotates to the right
ayddeg = ayd/pi*180
# rotation matrix of y-rotation
Ry = [[cos(ayd), 0, -sin(ayd)], [0, 1, 0], [sin(ayd), 0, cos(ayd)]]
# calculate x-translation of the car #######################################
# rotate Zc to let photo dibian parallel to XZplane, angle = aXZ
tl = y_fen / (ub-ut) # before rotation
al = atan(tl)
alz = al-aXZ # after rotation
# after y-rotation: Zc coordinates changes
Zcy = Ry*Zc
# if no x-translation: what is photo bottom line coordinates after rotation
vecbot0 = np.sign(ut-ub) * np.asarray([1, 0, (-Zcy[0][0]/Zcy[2][0])], dtype=float) # ut > ub: left lane
# what is photo middle line coordinates
vecmid0 = np.asarray([0, h, -(h*Zcy[1][0]/Zcy[2][0])], dtype=float)
vecbot0.shape = (1,3)
vecmid0.shape = (3,1)
# solve equation: dot product of vecl0 and vecl (vecl = vecmid0 - k*vecbot)
ak = np.power(np.linalg.norm(vecbot0, ord=2), 4)*(np.power(cos(alz),2)-1)
bk = 2*np.power(np.linalg.norm(vecbot0, ord=2), 2)*np.matmul(vecbot0,vecmid0)*(1-np.power(cos(alz), 2))
ck = np.power(np.linalg.norm(vecmid0, ord=2)*np.linalg.norm(vecbot0, ord=2)*cos(alz), 2) - np.power(np.matmul(vecbot0,vecmid0), 2)
bk, ck = bk[0][0], ck[0][0]
k1 = (-bk+sqrt(bk*bk - 4*ak*ck))/(2*ak)
k2 = (-bk-sqrt(bk*bk - 4*ak*ck))/(2*ak)
xd = max(k1,k2)
return xd,ayddeg
def draw_lines(img,lines,color=[255,0,0],thickness=3):
"""
划线
"""
if lines is None:
return
for line in lines:
for x1,y1,x2,y2 in line:
cv2.line(img,(x1,y1),(x2,y2),color,thickness)
return img
def group_lines_and_draw(img,lines,side):
"""
根据斜率,将所有的线分为左右两组,斜率绝对值小于0.5的舍去(减少噪声影响)
(因为图像的原点在左上角,slope<0是left lines,slope>0是right lines)
设定min_y作为left和right的top线,max_y为bottom线,求出四个x值即可确定直线:
将left和right的点分别线性拟合,拟合方程根据y值,求出x值,画出lane
"""
out_line = None
line_x,line_y=[],[]
for line in lines:
for x1,y1,x2,y2 in line:
slope=(y2-y1)/(x2-x1)
if side == 'Right':
if abs(slope)>0.4 and slope>0:
line_image=draw_lines(img,[[[x1,y1,x2,y2]]],color=[0,0,255],thickness=3)
line_x.extend([x1,x2])
line_y.extend([y1, y2])
# line_slope.extend([slope])
elif side == 'Left':
if abs(slope)>0.4 and slope<0:
line_image=draw_lines(img,[[[x1,y1,x2,y2]]],color=[0,0,255],thickness=3)
line_x.extend([x1,x2])
line_y.extend([y1, y2])
# line_slope.extend([slope])
#设定top 和 bottom的y值,y值都一样,根据ROI变化
min_y=int(-250)
max_y=int(img.shape[0]+90)
#对所有点进行线性拟合
if len(line_x)==0:
good = False
line_image = None
else:
poly_left = np.poly1d(np.polyfit(line_y, line_x, deg=1))
# Fit_slope = poly_left.coef[0]
# line_slope.extend([Fit_slope])
# var = np.var(line_slope)
# print('var',var)
x_start = int(poly_left(max_y))
x_end = int(poly_left(min_y))
out_line = [x_start,x_end]
good = True
line_image=draw_lines(img,[[
[x_start,max_y,x_end,min_y],
]],thickness=2)
return line_image,out_line,good
def Move_dicision(dx,Deg,Side='Right'):
if Side=='Right':
Distance_off = dx - 30
Deg_off = Deg
Step = abs(Distance_off) // 2
Trun = abs(Deg_off) // 5
if Distance_off > 0:
Move_action = 'Right02move'
else:
Move_action = 'Left02move'
if Deg_off < 0:
Turn_action = 'turn001R'
else:
Turn_action = 'turn001L'
# print('怎么走?' , Step, Trun, Move_action, Turn_action)
elif Side=='Left':
Distance_off = 30 - dx
Deg_off = Deg
Step = abs(Distance_off) // 2
Trun = abs(Deg_off) // 5
if Distance_off < 0:
Move_action = 'Left02move'
else:
Move_action = 'Right02move'
if Deg_off < 0:
Turn_action = 'turn001R'
else:
Turn_action = 'turn001L'
# print('怎么走?' , Step, Trun, Move_action, Turn_action)
return Step,Trun,Move_action,Turn_action
def Back_to_center (Chest_img,wich_side='Left'):
"""
split left and right to calculate
return list = Step,Trun,YouJinShen,Right,Clockwise
"""
Filter_length = 130
iteration = 0
while True:
if len(action_list) == 0:
print('Filter_length',Filter_length)
Chest = np.rot90(undistort_chest(Chest_img.imgs)).copy()
cv2.imshow("undistort_chest", Chest)
cv2.waitKey(1)
# continue
if wich_side == 'Right':
ROI_image = Chest[250:550,240:450]#右侧边缘,胸部
elif wich_side == 'Left':
ROI_image = Chest[250:550,30:239]#左侧边缘,胸部
# 机器人脚的位置
# ROI_image[340,:] = 255
cv2.imshow("Chest_img",ROI_image)
cv2.waitKey(1)
ROI_image = cv2.pyrMeanShiftFiltering(ROI_image, 9, 25)
cv2.imshow("pyrMeanShiftFiltering",ROI_image)
cv2.waitKey(1)
Canny_img = cv2.Canny(ROI_image,15,150)
# cv2.imshow("Canny_img",Canny_img)
# cv2.waitKey(1)
#膨胀加粗边缘
dilate = cv2.dilate(Canny_img, np.ones((2, 2), np.uint8), iterations=1)
cv2.imshow("dilate",dilate)
cv2.waitKey(1)
Lines = cv2.HoughLinesP(dilate,1.0,np.pi / 180, 100,minLineLength=Filter_length,maxLineGap=15)
# final_image = draw_lines(ROI_image,Lines,color=[0,255,0],thickness=2) #for test
# cv2.imshow("origine line",final_image)
# cv2.waitKey(1)
final_image, Final_line, good = group_lines_and_draw(ROI_image, Lines, wich_side)
if Final_line is None:
if Filter_length > 80:
Filter_length -= 10
else:
iteration += 1
continue
if iteration == 3:
print('No lines for long, just go')
break
cv2.imshow("image line",final_image)
cv2.waitKey(1)
# print('test')
if good:
if wich_side == 'Right':
Final_line[0] = Final_line[0] + 240
Final_line[1] = Final_line[1] + 240
if wich_side == 'Left':
Final_line[0] = Final_line[0] + 30
Final_line[1] = Final_line[1] + 30
dX, deg = Calculate_position(Final_line)
# print('line info',dX,deg)
Step, Trun, Move_action, Turn_action = Move_dicision(dX, deg, wich_side)
if Step == 0 and Trun == 0:
print('In the center')
break
else:
Step,Trun,Move_action,Turn_action = 0,0,True,True
print('啥也没看见朋友!')
for i in range(int(Trun)):
action_append(Turn_action)
time.sleep(0.5)
for i in range(int(Step)):
action_append(Move_action)
time.sleep(0.5)
def test_video(dataset):
while True:
cv2.imshow('img0', dataset.imgs)
if cv2.waitKey(1) == ord('q'): # q to quit
raise StopIteration
if __name__ == '__main__':
# Chest_path = '../code/track_picture/test/1005chest5.png'
# Chest_img = cv2.imread(Chest_path,1)
# Chest_img = cv2.resize(Chest_img, (640, 480))
Chest = LoadStreams(1)
# cv2.imshow("Chest_img",Chest_img)
# cv2.waitKey(1)
Back_to_center(Chest,'Left')