/
hsvthreshold_graph.py
283 lines (244 loc) · 7.27 KB
/
hsvthreshold_graph.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
import sys
sys.path.append('/users/sujit/libs/libfreenect/wrappers/python/')
import freenect
import math
import cv2
import numpy as np
import scipy.signal
import pylab
use_webcam = False
if use_webcam:
c = cv2.VideoCapture(0)
c.open(0)
plot_data = True
if plot_data:
pylab.ion()
graph, = pylab.plot([0], [0])
graph2, = pylab.plot([0], [0])
"""params = {
"huel": 0,
"hueh": 255,
"satl": 0,
"sath": 255,
"luml": 0,
"lumh": 255
}"""
#webcam with salmon paper
"""params = {
"hueh": 19,
"huel": 7,
"luml": 103,
"lumh": 255,
"sath": 206,
"satl": 71
}"""
#jpg with salmon paper
params = {
"hueh": 20,
"huel": 0,
"luml": 120,
"lumh": 255,
"sath": 255,
"satl": 0
}
"""params = {
"hueh": 74,
"huel": 52,
"luml": 10,
"lumh": 83,
"sath": 255,
"satl": 119
}"""
width = 640
height = 480
tolerance = 0.1
#imgpath = "wpitest.jpg"
#imgpath = "salmon3.jpg"
#imgpath = "testout.png"
imgpath = ""
comps = None
def getimg_webcam():
_, img = c.read()
return img
def getimg_irkinect():
raw_data, _ = freenect.sync_get_video(0, freenect.VIDEO_IR_8BIT)
return cv2.cvtColor(np.array(raw_data), cv2.COLOR_GRAY2BGR)
def getimg():
if use_webcam:
return getimg_webcam()
else:
return getimg_irkinect()
def donothing(img):
return img
def hsvthreshold(img):
hsv = cv2.cvtColor(img, cv2.COLOR_BGR2HSV)
lowerb = np.array([params['huel'], params['satl'], params['luml']])
upperb = np.array([params['hueh'], params['sath'], params['lumh']])
threshed = cv2.inRange(hsv, lowerb, upperb)
return threshed
def brightnessthreshold(img):
_, ret = cv2.threshold(img, 200, 255, cv2.THRESH_TOZERO)
return ret
def onchange(prop):
def ret(x):
params[prop] = x
cv2.setTrackbarPos(prop, 'processing', x)
return ret
def blurred(img):
return hsvthreshold(cv2.GaussianBlur(img, (21, 21), 1))
#return hsvthreshold(cv2.medianBlur(img, 21))
def contours(_img):
img = blurred(_img)
#img = hsvthreshold(_img)
ret = cv2.merge([img, img, img])
cons, hier = cv2.findContours(img, cv2.RETR_EXTERNAL, cv2.CHAIN_APPROX_NONE)
#cv2.drawContours(ret, np.array(cons), -1, np.array([255.0, 255.0, 0.0]))
for i in range(len(cons)):
if cv2.contourArea(cons[i]) >= 1000:
#cv2.drawContours(ret, np.array(cons), i, np.array([0.0, 0.0, 255.0]))
pts = []
for pt in cons[i]:
pts.append(pt[0])
ret[pt[0][1]][pt[0][0]] = np.array([255.0, 0.0, 0.0])
corn = corners(pts)
if len(corn) == 5 or True:
for x, y in corn:
cv2.circle(ret, (int(x), int(y)), 10, np.array([0.0, 0.0, 255.0]), -1)
"""corn = houghcorners(pts, img)
if corn != None:
for c in corn[0]:
x1 = int(c[0])
y1 = int(c[1])
x2 = int(c[2])
y2 = int(c[3])
cv2.line(ret, (x1, y1), (x2, y2), np.array([0.0, 0.0, 255.0]), 3)"""
return ret
def lderiv(pts):
#pts = [f(x - k), f(x - h), f(x)]
k = pts[2][0] - pts[0][0]
h = pts[1][0] - pts[0][0]
s = pts[1][1] - h ** 2 * pts[0][1] / k ** 2 + (h ** 2 / k ** 2 - 1) * pts[2][1]
return s / (h ** 2 / k - h)
def rderiv(pts):
#pts = [f(x), f(x + h), f(x + k)]
k = pts[2][0] - pts[0][0]
h = pts[1][0] - pts[0][0]
s = pts[1][1] - h ** 2 * pts[2][1] / k ** 2 + (h ** 2 / k ** 2 - 1) * pts[0][1]
return s / (h - h ** 2 / k)
def movingaverage(arr, window):
radius = (window - 1) / 2
weights = np.array([1.0] * window) / window
arr2 = np.concatenate([arr[-radius:], arr, arr[:radius]])
return np.convolve(arr2, weights, 'valid')
def houghcorners(pts, img):
img2 = np.zeros_like(img)
for pt in pts:
img2[pt[1]][pt[0]] = 1
segs = cv2.HoughLinesP(img2, 3, 0.001, len(pts) / 4 / 10, maxLineGap=len(pts))
return segs
def corners(pts):
pts = np.array(pts)
cx = pts[:,0].astype(np.double).sum() / len(pts)
cy = pts[:,1].astype(np.double).sum() / len(pts)
deltas = pts.astype(np.double) - [[cx, cy]] * len(pts)
#dists = movingaverage(np.sqrt(deltas[:,0] ** 2 + deltas[:,1] ** 2), 31)
dists = np.sqrt(deltas[:,0] ** 2 + deltas[:,1] ** 2)
angles = np.arctan2(deltas[:,0], deltas[:,1])
stacked = np.column_stack([deltas, dists, angles])
sorted = stacked[np.lexsort(stacked.transpose())]
ret = [[cx, cy]]
derivs = []
derivs2 = []
l = len(sorted)
prev = np.roll(sorted, -1, axis=0)
next = np.roll(sorted, 1, axis=0)
"""ld = (sorted[:,2] - prev[:,2]) / (sorted[:,3] - prev[:,3])
rd = (sorted[:,2] - next[:,2]) / (sorted[:,3] - next[:,3])
r = sorted[:,2]
theta = sorted[:,3]
s = np.sin(theta)
c = np.cos(theta)
d1 = (next[:,2] - prev[:,2]) / (next[:,3] - prev[:,3])"""
#derivs = np.abs(ld - rd)
#derivs = scipy.signal.medfilt(np.arctan2(d1 * s + sorted[:,1], d1 * c - sorted[:,0]), 1)
#xderiv = (next[:,1] - prev[:,1]) / (next[:,3] - prev[:,3])
#yderiv = (next[:,0] - prev[:,0]) / (next[:,3] - prev[:,3])
nnext = np.roll(sorted, -10, axis=0)
dx = nnext[:,1] - sorted[:,1]
dy = nnext[:,0] - sorted[:,0]
window = len(pts) / 8
if window % 2 == 0:
window += 1
slopes = np.arctan2(dy, dx)
derivs = scipy.signal.medfilt(slopes, window)
"""pderiv = np.roll(derivs, -1)
nderiv = np.roll(derivs, 1)
d2 = np.abs((nderiv - pderiv) / (next[:,3] - prev[:,3]))"""
if plot_data:
xdata = sorted[:,3]
ydata = sorted[:,2]
ydata2 = scipy.signal.medfilt(slopes, 21) / np.abs(slopes).max() * np.abs(ydata).max()
graph.set_xdata(xdata)
graph.set_ydata(ydata)
graph.axes.set_xlim(xdata.min(), xdata.max())
graph.axes.set_ylim(ydata.min(), ydata.max())
graph2.set_xdata(xdata)
graph2.set_ydata(ydata2)
graph2.axes.set_xlim(xdata.min(), xdata.max())
graph2.axes.set_ylim(ydata2.min(), ydata2.max())
pylab.draw()
for i in range(l):
perfect = True
"""prev = (i - 30) % l
pprev = (i - 2) % l
next = (i + 30) % l
nnext = (i + 2) % l
#ld = lderiv([sorted[pprev][3:1:-1], sorted[prev][3:1:-1], sorted[i][3:1:-1]])
#rd = rderiv([sorted[i][3:1:-1], sorted[next][3:1:-1], sorted[nnext][3:1:-1]])
ld = (sorted[i][2] - sorted[prev][2]) / (sorted[i][3] - sorted[prev][3])
rd = (sorted[i][2] - sorted[next][2]) / (sorted[i][3] - sorted[next][3])"""
#print "%f\t%f\t%f\t%f\t%f" % (sorted[i][3], sorted[i][2], slopes[i], derivs[i], d2[i])
#print "%f\t%f\t%f" % (sorted[i][3], slopes[i], derivs[i])
#print "%d\t%d" % (pts[i][0], pts[i][1])
for k in range(1, 10):
next = (i + k) % l
prev = (i - k) % l
#if not (d2[i] > d2[next] and d2[i] > d2[prev]):
if not (sorted[i][2] > sorted[next][2] and sorted[i][2] > sorted[prev][2]):
perfect = False
break
#print "%f, %f, %f" % (sorted[prev][2], sorted[i][2], sorted[next][2])"""
if perfect:
ret.append(sorted[i][:2] + [cx, cy])
"""for quad in range(4):
best = quad * l / 4
for i in range(quad * l / 4 + 1, (quad + 1) * l / 4):
if sorted[i][2] > sorted[best][2]:
best = i
ret.append(sorted[best][:2] + [cx, cy])"""
return ret
def loop(processimg):
#if not use_webcam and imgpath == "":
# ctx = freenect.init()
# dev = freenect.open_device(ctx, 0)
# freenect.set_tilt_degs(dev, 10)
# freenect.close_device(dev)
cv2.namedWindow('processing')
for k, v in params.iteritems():
cv2.createTrackbar(k, 'processing', v, 255, onchange(k))
runonce = True
while runonce:
#runonce = False
if imgpath != "":
img = cv2.imread(imgpath)
else:
img = getimg()
cv2.imshow('processing', cv2.resize(processimg(img), (width, height)))
char = cv2.waitKey(10)
if char == 27:
break
elif char == ord('p'):
for k, v in params.iteritems():
print "%s: %d" % (k, v)
cv2.destroyAllWindows()
loop(contours)