import matplotlib as mpl
mpl.use('Agg')
import picamera, cv2
from lib.utils import dbprint, html_data_path
from lib.camera import update_img, ColorFix
clrLower = [0, 10, 10]
clrUpper = [64, 255, 255]
if __name__ == '__main__':
with picamera.PiCamera() as camera:
ss = ColorFix(camera)
try:
# data = ss.mask_range(clrLower, clrUpper)
data = ss.blob_detection()
if data:
# cv2.imwrite(html_data_path('frame_%03d.jpg' % i), data['frame'])
# cv2.imwrite(html_data_path('iframe_%03d.jpg' % i), data['iframe'])
# cv2.imwrite(html_data_path('oframe_%03d.jpg' % i), data['oframe'])
cv2.imwrite(html_data_path('frame.jpg'), data['frame'])
cv2.imwrite(html_data_path('iframe.jpg'), data['iframe'])
cv2.imwrite(html_data_path('oframe.jpg'), data['oframe'])
# json.dump(data['hlist'], file('colorfix.json', 'w'))
else:
dbprint("NOT FOUND")
# json.dump({'imgcount': i}, file(html_data_path('frames.json'), 'w'), indent=2)
finally:
update_img(camera)
from lib.utils import html_data_path
from lib.utils import dbprint
from lib.camera import update_img, FeatureProcess
from lib.frobo_ng import frobo_ng
if __name__ == '__main__':
c = frobo_ng()
#c.debug = True
with picamera.PiCamera() as camera:
fp = FeatureProcess(camera)
try:
update_img(camera, html_data_path('pic0.jpg'))
c.turn(350)
time.sleep(1)
update_img(camera, 'images/pic1.jpg'))
fp.percent()
c.move_straight(fwd=True, max_steps=c.m2steps(0.5), max_secs=1)
update_img(camera, html_data_path('pic2.jpg'))
time.sleep(1)
c.turn(90)
time.sleep(1)
c.move_straight(fwd=True, max_steps=c.m2steps(0.5), max_secs=1)
update_img(camera, html_data_path('pic3.jpg'))
time.sleep(1)
c.turn(210)
time.sleep(1)
c.move_straight(fwd=True, max_steps=c.m2steps(0.5), max_secs=1)
#!/usr/bin/env python
import sys, os, time, json
import picamera, cv2
from lib.utils import dbprint, html_data_path
from lib.camera import update_img, ShapeSearch, capture_cvimage
if __name__ == '__main__':
with picamera.PiCamera() as camera:
ss = ShapeSearch(camera)
try:
data = ss.find_shapes(threshold=127)
if data:
if os.environ.get('RASPICAM_ROTATE', ''):
angle = int(os.environ['RASPICAM_ROTATE'])
rows, cols, depth = data['frame'].shape
M = cv2.getRotationMatrix2D((cols / 2, rows / 2), 180, 1)
data['frame'] = cv2.warpAffine(data['frame'], M,
(cols, rows))
cv2.imwrite(html_data_path('shapes.jpg'), data['frame'])
cv2.imwrite(html_data_path('thresh.jpg'), data['thresh'])
finally:
update_img(camera)
fy = sp.interp1d(tlist, ylist, kind='linear')
tlist = np.linspace(min(tlist), max(tlist), 50)
dots[k] = []
for i in range(len(tlist)):
dots[k].append({'t': tlist[i], 'x': fx(tlist[i]), 'y': fy(tlist[i])})
im = Image.new('RGBA', (800, 800), (240, 240, 240, 0))
draw = ImageDraw.Draw(im)
colors = ('red', 'blue')
ci = 0
zoom = 4
for k in dots.keys():
if dots[k]:
clr = colors[ci]
ci += 1
x = dots[k][0]['x'] * zoom
y = dots[k][0]['y'] * zoom
draw.ellipse((x - 5, y - 5, x + 5, y + 5), fill=clr, outline=clr)
for d in dots[k][1:]:
draw.line((x, y, d['x'] * zoom, d['y'] * zoom), fill=clr)
x = d['x'] * zoom
y = d['y'] * zoom
# a_x = d['acc']['x']
# a_y = d['acc']['y']
# draw.line((x, y, x + a_x, y + a_y), fill='green')
# if d['dist'] > 0 and d['dist'] < 20:
# draw.ellipse((x-2, y-2, x+2, y+2), fill='orange', outline='orange')
# if d.get('hit_warn', None):
# draw.ellipse((x-4, y-4, x+4, y+4), fill='yellow', outline='red')
im.save(html_data_path('drawing.jpg'))
cnt = 20
for tt in [
cv2.THRESH_BINARY, cv2.THRESH_BINARY_INV, cv2.THRESH_TRUNC,
cv2.THRESH_TOZERO, cv2.THRESH_TOZERO_INV
]:
for t in range(t_start, 255, (t_end - t_start) / cnt):
data = ss.find_shapes(threshold=t, threshold_type=tt)
if data:
if os.environ.get('RASPICAM_ROTATE', ''):
angle = int(os.environ['RASPICAM_ROTATE'])
rows, cols, depth = data['frame'].shape
M = cv2.getRotationMatrix2D((cols / 2, rows / 2),
180, 1)
data['frame'] = cv2.warpAffine(
data['frame'], M, (cols, rows))
cv2.imwrite(html_data_path('shapes_%03d.jpg' % i),
data['frame'])
# cv2.putText(data['thresh'], 'Thresh: %d' % t, (10, 20), cv2.FONT_HERSHEY_SIMPLEX, 1, (255, 0, 255), 2)
cv2.putText(data['thresh'],
'Thresh: %d, Type: %d' % (t, tt), (10, 20),
cv2.FONT_HERSHEY_SIMPLEX, 0.6,
(255, 255, 255), 2)
cv2.imwrite(html_data_path('thresh_%03d.jpg' % i),
data['thresh'])
i += 1
dbprint(('*' * 10) + ' Image %d written' % i)
json.dump({'imgcount': i},
file(html_data_path('shapes.json'), 'w'),
indent=2)
finally:
update_img(camera)
import picamera, cv2
from lib.utils import dbprint, html_data_path
from lib.camera import update_img, FeatureProcess, capture_cvimage
from lib.frobo_ng import frobo_ng
if __name__ == '__main__':
c = frobo_ng()
# c.debug = True
with picamera.PiCamera() as camera:
fp = FeatureProcess(camera)
try:
fp.percent()
c.tick_left(min_angle=10)
data = fp.percent()
if data:
cv2.imwrite(html_data_path('pic1.jpg'), data['frame'])
dbprint('Left=%g' % (data['percent'] if data else data))
cv2.imwrite(html_data_path('frame.jpg'), data['frame'])
c.tick_right(min_angle=20)
data = fp.percent()
if data:
cv2.imwrite(html_data_path('pic2.jpg'), data['frame'])
dbprint('Right=%g' % (data['percent'] if data else data))
finally:
update_img(camera)
#!/usr/bin/env python
import sys, os, time, json
import picamera, cv2
from lib.utils import dbprint, html_data_path
from lib.camera import update_img, StereoDisparity
from lib.frobo_ng import frobo_ng
if __name__ == '__main__':
c = frobo_ng()
# c.debug = True
with picamera.PiCamera() as camera:
fp = StereoDisparity(camera)
try:
c.tick_left(min_angle=3)
fp.left_frame()
c.tick_right(min_angle=3)
fp.right_frame()
fp.write_ply(html_data_path('disparity.ply'))
cv2.imwrite(html_data_path('pic0.jpg'), fp.lframe)
cv2.imwrite(html_data_path('pic1.jpg'), fp.rframe)
cv2.imwrite(html_data_path('pic2.jpg'), fp.disparity)
finally:
update_img(camera)
#D = 808 * 0.03 / 50.68 ~ 0.478 (measured 0.49)
#D = 808 * 0.03 / 32.31 ~ 0.75 (measured 0.745)
if __name__ == '__main__':
if 1:
r = redis.Redis()
use_camera(r, width=1280, height=960)
# use_camera(r, brightness=80, contrast=85)
# use_camera(r, width=640, height=480, brightness=80, contrast=80)
# use_camera(r, brightness=90, contrast=90)
time.sleep(4)
try:
pass
# markers = collect_markers(r, fpath = html_data_path('markers.jpg'))
# print json.dumps(dict([(m['id'], m['distance']) for m in markers]), indent=2)
finally:
time.sleep(1)
make_shot(r, fpath='redis:image')
rimg = r.get('image')
udp_send('opcplus', data_name='hubee.jpg', data=rimg)
imgdata = np.asarray(bytearray(rimg), dtype=np.uint8)
img = cv2.imdecode(imgdata, cv2.CV_LOAD_IMAGE_COLOR)
cv2.imwrite(html_data_path('picam_0.jpg'), img)
release_camera(r)
else:
cam = picamera.PiCamera()
time.sleep(2)
# update_img(cam, brightness=90, contrast=90, exposure_mode='off')
update_img(cam)
def clear_img():
unlink(html_data_path('start.jpg'))
for f in glob.glob(html_data_path('found*.jpg')):
unlink(f)
for f in glob.glob(html_data_path('circle*.jpg')):
unlink(f)
def update_img(camera):
camera.exposure_mode = 'off'
camera.brightness = 50
camera.contrast = 50
camera.capture(html_data_path('picam_0.jpg'), use_video_port=True)
def experiment2(c, camera):
def filterMatches(base_kp, kp, matches, ratio=0.75):
mkp1, mkp2 = [], []
for m in matches:
if len(m) == 2 and m[0].distance < m[1].distance * ratio:
m = m[0]
mkp1.append(base_kp[m.queryIdx])
mkp2.append(kp[m.trainIdx])
return zip(mkp1, mkp2)
try:
clear_img()
camera.resolution = (160, 120)
camera.framerate = 5
cv_det = cv2.FeatureDetector_create(DETECTOR)
cv_desc = cv2.DescriptorExtractor_create(EXTRACTOR)
matcher = cv2.DescriptorMatcher_create(MATCHER)
was_below = False
compass = hmc5883l(gauss=4.7, declination=(12, 34))
old_heading = init_heading = compass.heading()
hbase_diff = 0
h_dir = 1
with picamera.array.PiRGBArray(camera) as stream:
camera.capture(stream, format='bgr', use_video_port=True)
base_frame = stream.array
cv2.imwrite(html_data_path('start.jpg'), base_frame)
circle_count = 0
found_count = 0
base_kp = cv_det.detect(base_frame)
base_kp, base_desc = cv_desc.compute(base_frame, base_kp)
base_kpl = len(base_kp)
for i in range(200):
c.right_move(True, 40)
c.left_move(False, 40)
time.sleep(0.1)
c.stop()
time.sleep(0.2)
with picamera.array.PiRGBArray(camera) as stream:
camera.capture(stream, format='bgr', use_video_port=True)
frame = stream.array
kp = cv_det.detect(frame)
kp, desc = cv_desc.compute(frame, kp)
matches = matcher.knnMatch(base_desc,
trainDescriptors=desc,
k=2)
pairs = filterMatches(base_kp, kp, matches)
lp = len(pairs)
rperc = (lp * 100) / base_kpl
heading = compass.heading()
hdiff = abs(heading - init_heading)
if h_dir > 0:
if hbase_diff + 5 <= hdiff:
hbase_diff = hdiff
elif hbase_diff >= hdiff + 5:
h_dir = -1
else:
if hbase_diff >= hdiff + 5:
hbase_diff = hdiff
elif hbase_diff + 5 <= hdiff:
# found
dbprint('Circle!!!')
circle_count += 1
cv2.imwrite(
html_data_path('circle%03d.jpg' % circle_count),
frame)
h_dir = 1
dbprint('%.2f%% - %.2f (%.2f, dh: %.2f)' %
(rperc, heading, hdiff, abs(old_heading - heading)))
if was_below:
if rperc > 60:
dbprint('Found close')
was_below = False
found_count += 1
cv2.imwrite(
html_data_path('found%03d.jpg' % found_count),
frame)
else:
if rperc < 10:
was_below = True
old_heading = heading
finally:
c.stop()
dbprint('stopped')
from lib.frobo_ng import frobo_ng
AZIM_SHOT = 350
AZIM_MOVE = 90
if __name__ == '__main__':
c = frobo_ng()
c.debug = True
with picamera.PiCamera() as camera:
fp = FeatureProcess(camera)
try:
update_img(camera, html_data_path('pic0.jpg'))
c.turn(AZIM_SHOT)
update_img(camera, html_data_path('pic1.jpg'))
fp.percent()
c.turn(AZIM_MOVE)
c.move_straight(fwd=True, max_secs=1, max_steps=100)
c.turn(AZIM_SHOT)
dbprint('Matches %s%%' % fp.percent())
update_img(camera, html_data_path('pic2.jpg'))
c.turn(AZIM_MOVE)
c.move_straight(fwd=True, max_secs=1, max_steps=100)
c.turn(AZIM_SHOT)
dbprint('Matches %s%%' % fp.percent())
update_img(camera, html_data_path('pic3.jpg'))
c.turn(AZIM_MOVE)
c.move_straight(fwd=True, max_secs=1, max_steps=100)