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
|
import cv2
import math
import numpy
from math import sin,cos,tan,sqrt
cap = cv2.VideoCapture("../vid.mp4")
writer = cv2.VideoWriter("../outvid.avi",cv2.cv.CV_FOURCC(*'MP42'),25,(1600,600),1)
###### CONFIG ######
feature_params = dict( maxCorners = 100,
qualityLevel = 0.3,
minDistance = 20,
blockSize = 7 )
scale_factor=0.5
scr_width=1600
scr_height=600
# Parameters for lucas kanade optical flow
lk_params = dict( winSize = (15,15),
maxLevel = 2,
criteria = (cv2.TERM_CRITERIA_EPS | cv2.TERM_CRITERIA_COUNT, 10, 0.03))
###### INITALISATION ######
def get_maps(xres,yres):
xlist=[]
ylist=[]
constant = 30./41.5*3
# calculated from official spec: maximum angle (i.e., over the diagonals)
# is 92 deg. -> angle over half a diagonal's length is 92/2 deg
constant2= 1./(math.sqrt((1280/2)**2+(720/2)**2))*(92/2)/180*math.pi
foo = math.tan( 92/2/math.sqrt(16**2+9**2)*16. ) * constant / math.tan(1280/2 * constant2)
d = 18. # distance from center to camera
r = 200. # distance from center to recorded objects
alpha = 92/2/math.sqrt(16**2+9**2)*16. # angle over x axis
lam = -d * cos(alpha) + math.sqrt(d**2 * (cos(alpha))**2 - d**2 + r**2)
dist_canvas = d + lam*cos(alpha)
width_canvas = lam*sin(alpha) * 2
beta = math.atan(width_canvas/2./dist_canvas) # viewing angle as calculated from center of camera. depends from object distance, because camera is not at the center.
for y in xrange(0,yres):
xtmp=[]
ytmp=[]
yy=(y-yres/2.)/xres # yes, xres.
for x in xrange(0,xres):
#xx= tan((x-xres/2.)/xres*2.*beta) * dist_canvas / width_canvas
#xx= tan((x-xres/2.)/xres*width_canvas/dist_canvas) * dist_canvas / width_canvas
xx= tan((x-xres/2.)/xres*2.*beta) /2./beta
# derivative of this at location 0 should be 1.0/xres
dist = math.sqrt(xx**2 + yy**2)
if (dist != 0):
xtmp=xtmp+[ 1280/2+ xx/dist/constant2*math.atan(constant*dist) ]
ytmp=ytmp+[ 720/2 + yy/dist/constant2*math.atan(constant*dist) ]
else:
xtmp=xtmp+[0+1280/2]
ytmp=ytmp+[0+720/2]
xlist=xlist+[xtmp]
ylist=ylist+[ytmp]
if y % 10 == 0:
print y
xmap=numpy.array(xlist).astype('float32')
ymap=numpy.array(ylist).astype('float32')
return xmap,ymap
xmap,ymap = get_maps(1280,720)
ret,oldframe_=cap.read()
rawheight, rawwidth, bpp = oldframe_.shape
oldframe = cv2.remap(oldframe_, xmap, ymap, cv2.INTER_CUBIC)
oldgray=cv2.cvtColor(oldframe,cv2.COLOR_BGR2GRAY)
height, width, bpp = oldframe.shape
mask_ = numpy.ones((rawheight,rawwidth, 1), numpy.uint8) * 255
mask = cv2.remap(mask_, xmap, ymap, cv2.INTER_CUBIC)
screencontent = numpy.zeros((scr_height, scr_width,3), numpy.uint8)
total_angle=0.
total_x=1500
total_y=-300
while(cap.isOpened()):
ret, frame_ = cap.read()
frame = cv2.remap(frame_, xmap, ymap, cv2.INTER_CUBIC)
gray = cv2.cvtColor(frame,cv2.COLOR_BGR2GRAY)
# calculate movement, rotation and stretch. (we will ignore the stretch factor.)
mat = cv2.estimateRigidTransform(gray,oldgray,False)
angle = math.atan2(mat[0,1],mat[0,0])
stretch = int((math.sqrt(mat[0,1]**2+mat[0,0]**2)-1)*100)
# calculate shift_x and _y is if one would rotate-and-stretch around the center of the image, not the topleft corner
shift_x = mat[0,2] - width/2 + ( mat[0,0]*width/2 + mat[0,1]*height/2 )
shift_y = mat[1,2] - height/2 + ( mat[1,0]*width/2 + mat[1,1]*height/2 )
# accumulate values
total_x = total_x + shift_x
total_y = total_y + shift_y
total_angle=total_angle+angle
print angle/3.141592654*180,'deg\t',stretch,"%\t", shift_x,'\t',shift_y
# rotate and move current frame into a global context
mat2=cv2.getRotationMatrix2D((width/2,height/2), total_angle/3.141593654*180, scale_factor)
mat2[0,2] = mat2[0,2]+total_x*scale_factor
mat2[1,2] = mat2[1,2]+total_y*scale_factor
frame2= cv2.warpAffine(frame, mat2, (scr_width,scr_height) )
mask2 = cv2.warpAffine(mask, mat2, (scr_width,scr_height) )
ret, mask2 = cv2.threshold(mask2, 254, 255, cv2.THRESH_BINARY)
mask3 = cv2.erode(mask2, numpy.ones((20,200),numpy.uint8)) # strip off the potentially-badlooking edges. left/right borders are darkened, strip them off
mask4 = cv2.erode(mask2, numpy.ones((1,1),numpy.uint8))
screencontent = cv2.bitwise_and(screencontent,screencontent, mask=cv2.bitwise_not(mask3)) # blank out
screencontent = cv2.add(screencontent, cv2.bitwise_and(frame2,frame2,mask=mask3)) # and redraw
screencontent2=screencontent.copy()
screencontent2=cv2.bitwise_and(screencontent2,screencontent2, mask=cv2.bitwise_not(mask4))
screencontent2=cv2.add(screencontent2, cv2.bitwise_and(frame2,frame2,mask=mask4))
cv2.imshow('frame', frame)
cv2.imshow('screencontent', screencontent)
cv2.imshow('screencontent2', screencontent2)
writer.write(screencontent2)
oldframe=frame
oldgray=gray
key = cv2.waitKey(1) & 0xFF
if key == ord("q"):
break
if key == ord("r"):
total_angle=0.
total_x=700
total_y=-200
screencontent = numpy.zeros((scr_height, scr_width,3), numpy.uint8)
|