def transformFrame(img, twoRange, strict_ranges, loose_ranges):
img = img.toHSV()
res = cv2.inRange(img.getNumpyCv2(), np.array(strict_ranges[0]), np.array(strict_ranges[1]))
simg = scv.Image(res.transpose(1, 0))
if twoRange:
blobs = simg.findBlobs()
if blobs:
x = int(blobs[0].minRectX() - 20)
if x < 0:
x = 0
y = int(blobs[0].minRectY() - 20)
if y < 0:
y = 0
w = int(blobs[0].minRectWidth() + 40)
h = int(blobs[0].minRectHeight() + 40)
bb = (x, y, w, h)
cropped = img.crop(bb)
loose_res = cv2.inRange(cropped.getNumpyCv2(), np.array(loose_ranges[0]), np.array(loose_ranges[1]))
for i in range(y, y+cropped.height):
for j in range(x, x+cropped.width):
temp = loose_res[i - y, j - x]
res[i, j] = temp
simg = scv.Image(res.transpose(1, 0))
simg = simg.morphClose()
simg = simg.morphOpen()
return simg
def update(self, dt): self.img = GameInstance().cam.getImage().flipHorizontal() GameInstance().lower = np.array((float(self.val[0]), float(self.val[1]), float(self.val[2]))) GameInstance().upper = np.array((float(self.val[3]), float(self.val[4]), float(self.val[5]))) hsv = self.img.toHSV() mask = cv2.inRange(hsv.getNumpy(), GameInstance().lower, GameInstance().upper) img2 = Image(source=mask) img2 = img2.erode(1) img2 = img2.dilate(2) if self.cvActiveBloops.active: blops = img2.findBlobs() if blops: self.cvBlob.texture = getKivyTexture(blops[-1].getMaskedImage()) self.cvFilter.texture = getKivyTexture(img2) self.cvImage.texture = getKivyTexture(self.img)
def update(self, dt):
self.img = GameInstance().cam.getImage().flipHorizontal()
GameInstance().lower = np.array(
(float(self.val[0]), float(self.val[1]), float(self.val[2])))
GameInstance().upper = np.array(
(float(self.val[3]), float(self.val[4]), float(self.val[5])))
hsv = self.img.toHSV()
mask = cv2.inRange(hsv.getNumpy(),
GameInstance().lower,
GameInstance().upper)
img2 = Image(source=mask)
img2 = img2.erode(1)
img2 = img2.dilate(2)
if self.cvActiveBloops.active:
blops = img2.findBlobs()
if blops:
self.cvBlob.texture = getKivyTexture(
blops[-1].getMaskedImage())
self.cvFilter.texture = getKivyTexture(img2)
self.cvImage.texture = getKivyTexture(self.img)
def _initData(self):
"""
Initialize the cluster centers and initial values of the pixel-wise
cluster assignment and distance values.
"""
self.clusters = -1 * np.ones(self.img.shape[:2])
self.distances = self.FLT_MAX * np.ones(self.img.shape[:2])
centers = []
for i in xrange(self.step, self.width - self.step/2, self.step):
for j in xrange(self.step, self.height - self.step/2, self.step):
nc = self._findLocalMinimum(center=(i, j))
color = self.labimg[nc[1], nc[0]]
center = [color[0], color[1], color[2], nc[0], nc[1]]
centers.append(center)
self.center_counts = np.zeros(len(centers))
self.centers = np.array(centers)
def predictY(self):
"""
**SUMMARY**
Returns a numpy array of the predicted y (vertical) coordinate of each feature.
**RETURNS**
A numpy array.
**EXAMPLE**
>>> while True:
... img1 = cam.getImage()
... ts = img1.track("camshift", ts1, img, bb)
... img = img1
>>> print ts.predictY()
"""
return np.array([f.predict_pt[1] for f in self])
def correctedCoordinates(self):
"""
**SUMMARY**
Returns a numpy array of the corrected coordinates of each feature.
**RETURNS**
A numpy array.
**EXAMPLE**
>>> while True:
... img1 = cam.getImage()
... ts = img1.track("camshift", ts1, img, bb)
... img = img1
>>> print ts.predictedCoordinates()
"""
return np.array([f.state_pt for f in self])
def predictY(self):
"""
**SUMMARY**
Returns a numpy array of the predicted y (vertical) coordinate of each feature.
**RETURNS**
A numpy array.
**EXAMPLE**
>>> while True:
... img1 = cam.getImage()
... ts = img1.track("camshift", ts1, img, bb)
... img = img1
>>> print ts.predictY()
"""
return np.array([f.predict_pt[1] for f in self])
def correctedCoordinates(self):
"""
**SUMMARY**
Returns a numpy array of the corrected coordinates of each feature.
**RETURNS**
A numpy array.
**EXAMPLE**
>>> while True:
... img1 = cam.getImage()
... ts = img1.track("camshift", ts1, img, bb)
... img = img1
>>> print ts.predictedCoordinates()
"""
return np.array([f.state_pt for f in self])
def pixleVelocityRealTime(self):
"""
**SUMMARY**
Get each Pixel Velocity of the tracked object in pixel/frames.
**PARAMETERS**
No Parameters required.
**RETURNS**
* *numpy array* * - array of pixel velocity tuple.
>>> while True:
... img1 = cam.getImage()
... ts = img1.track("camshift", ts1, img, bb)
... img = img1
>>> print ts.pixelVelocityRealTime()
"""
return np.array([f.rt_vel for f in self])
def pixleVelocityRealTime(self):
"""
**SUMMARY**
Get each Pixel Velocity of the tracked object in pixel/frames.
**PARAMETERS**
No Parameters required.
**RETURNS**
* *numpy array* * - array of pixel velocity tuple.
>>> while True:
... img1 = cam.getImage()
... ts = img1.track("camshift", ts1, img, bb)
... img = img1
>>> print ts.pixelVelocityRealTime()
"""
return np.array([f.rt_vel for f in self])
def areaRatio(self):
"""
**SUMMARY**
Returns a numpy array of the areaRatio of each feature.
where areaRatio is the ratio of the size of the current bounding box to
the size of the initial bounding box
**RETURNS**
A numpy array.
**EXAMPLE**
>>> while True:
... img1 = cam.getImage()
... ts = img1.track("camshift", ts1, img, bb)
... img = img1
>>> print ts.areaRatio
"""
return np.array([f.areaRatio for f in self])
def areaRatio(self):
"""
**SUMMARY**
Returns a numpy array of the areaRatio of each feature.
where areaRatio is the ratio of the size of the current bounding box to
the size of the initial bounding box
**RETURNS**
A numpy array.
**EXAMPLE**
>>> while True:
... img1 = cam.getImage()
... ts = img1.track("camshift", ts1, img, bb)
... img = img1
>>> print ts.areaRatio
"""
return np.array([f.areaRatio for f in self])
from trash.vision.scribbles import Trackbars
tbars = Trackbars()
ranges = tbars.getFromUser()
video = scv.Camera()
display = scv.Display()
while display.isNotDone():
img = video.getImage()
img = img.toHSV()
res = cv2.inRange(img.getNumpyCv2(), np.array(ranges[0]), np.array(ranges[1]))
#simg = scv.Image(thres, cv2image=False)
simg = scv.Image(res.transpose(1,0))
#simg = simg.morphOpen()
simg = simg.morphClose()
simg = simg.morphOpen()
#znajdz najwiekszego bloba na masce i on pewnie bedzie pileczka, nowa maska
#ewentualnie podejscie z dwoma filtrami, jeden "surowy", ma znalezc tylko kawalek pilki
# a drugi lagodniejszy, ale dolaczamy tlyko to, co przylega do znalezionych
# powinno rozrosnac znaleziony fragment pilki do calej pilki
blobs = simg.findBlobs()
if blobs:
class __GameInstance:
lower = np.array((float(0), float(0), float(0)))
upper = np.array((float(0), float(0), float(0)))
def __init__(self):
self.cam = Camera()
([34, 25, 51], [38, 255, 255]),
([72, 0, 137], [85, 23, 182]),
([45, 3, 110], [67, 31, 162]),
([21, 23, 0], [49, 237, 67])
]
display = scv.Display()
while display.isNotDone():
img = video.getImage()
img = img.toHSV()
thres = []
for range in ranges:
thres.append(cv2.inRange(img.getNumpyCv2(), np.array(range[0]), np.array(range[1])))
res = reduce(lambda x, y: cv2.add(x, y), thres)
#simg = scv.Image(thres, cv2image=False)
simg = scv.Image(res.transpose(1,0))
#simg = simg.morphOpen()
simg = simg.morphClose()
simg = simg.morphOpen()
#blobs = simg.findBlobs()
#if blobs:
# circles = blobs.filter([b.isCircle(0.4) for b in blobs])
# if circles:
# # for c in circles:
display = scv.Display()
history = []
tracker = Tracker()
while display.isNotDone():
img = video.getImage()
img = img.toHSV()
#apply strict range, find biggest blob
#near the biggest blob apply loose range
res = cv2.inRange(img.getNumpyCv2(), np.array(strict_ranges[0]), np.array(strict_ranges[1]))
simg = scv.Image(res.transpose(1, 0))
blobs = simg.findBlobs()
if blobs:
x = int(blobs[0].minRectX() - 20)
if x < 0:
x = 0
y = int(blobs[0].minRectY() - 20)
if y < 0:
y = 0
w = int(blobs[0].minRectWidth() + 40)
h = int(blobs[0].minRectHeight() + 40)
bb = (x, y, w, h)
# print x,y,w,h
def camshiftTracker(img, bb, ts, **kwargs):
"""
**DESCRIPTION**
(Dev Zone)
Tracking the object surrounded by the bounding box in the given
image using CAMshift method.
Warning: Use this if you know what you are doing. Better have a
look at Image.track()
**PARAMETERS**
* *img* - Image - Image to be tracked.
* *bb* - tuple - Bounding Box tuple (x, y, w, h)
* *ts* - TrackSet - SimpleCV.Features.TrackSet.
Optional PARAMETERS:
lower - Lower HSV value for inRange thresholding
tuple of (H, S, V)
upper - Upper HSV value for inRange thresholding
tuple of (H, S, V)
mask - Mask to calculate Histogram. It's better
if you don't provide one.
num_frames - number of frames to be backtracked.
**RETURNS**
SimpleCV.Features.Tracking.CAMShift
**HOW TO USE**
>>> cam = Camera()
>>> ts = []
>>> img = cam.getImage()
>>> bb = (100, 100, 300, 300) # get BB from somewhere
>>> ts = CAMShiftTracker(img, bb, ts, lower=(40, 120, 120), upper=(80, 200, 200), num_frames=30)
>>> while (some_condition_here):
... img = cam.getImage()
... bb = ts[-1].bb
... ts = CAMShiftTracker(img, bb, ts, lower=(40, 120, 120), upper=(80, 200, 200), num_frames=30)
... ts[-1].drawBB()
... img.show()
This is too much confusing. Better use
Image.track() method.
READ MORE:
CAMShift Tracker:
Uses meanshift based CAMShift thresholding technique. Blobs and objects with
single tone or tracked very efficiently. CAMshift should be preferred if you
are trying to track faces. It is optimized to track faces.
"""
lower = np.array((0., 60., 32.))
upper = np.array((180., 255., 255.))
mask = None
num_frames = 40
if not isinstance(bb, tuple):
bb = tuple(bb)
bb = (int(bb[0]), int(bb[1]), int(bb[2]), int(bb[3]))
for key in kwargs:
if key == 'lower':
lower = np.array(tuple(kwargs[key]))
elif key == 'upper':
upper = np.array(tuple(kwargs[key]))
elif key == 'mask':
mask = kwargs[key]
mask = mask.getNumpyCv2()
elif key == 'num_frames':
num_frames = kwargs[key]
hsv = cv2.cvtColor(img.getNumpyCv2(), cv2.cv.CV_BGR2HSV)
if mask is None:
mask = cv2.inRange(hsv, lower, upper)
x0, y0, w, h = bb
x1 = x0 + w - 1
y1 = y0 + h - 1
hsv_roi = hsv[y0:y1, x0:x1]
mask_roi = mask[y0:y1, x0:x1]
hist = cv2.calcHist([hsv_roi], [0], mask_roi, [16], [0, 180])
cv2.normalize(hist, hist, 0, 255, cv2.NORM_MINMAX)
hist_flat = hist.reshape(-1)
imgs = [hsv]
if len(ts) > num_frames and num_frames > 1:
for feat in ts[-num_frames:]:
imgs.append(feat.image.toHSV().getNumpyCv2())
elif len(ts) < num_frames and num_frames > 1:
for feat in ts:
imgs.append(feat.image.toHSV().getNumpyCv2())
prob = cv2.calcBackProject(imgs, [0], hist_flat, [0, 180], 1)
prob &= mask
term_crit = (cv2.TERM_CRITERIA_EPS | cv2.TERM_CRITERIA_COUNT, 10, 1)
new_ellipse, track_window = cv2.CamShift(prob, bb, term_crit)
if track_window[2] == 0 or track_window[3] == 0:
track_window = bb
track = CAMShiftTrack(img, track_window, new_ellipse)
return track
def camshiftTracker(img, bb, ts, **kwargs):
"""
**DESCRIPTION**
(Dev Zone)
Tracking the object surrounded by the bounding box in the given
image using CAMshift method.
Warning: Use this if you know what you are doing. Better have a
look at Image.track()
**PARAMETERS**
* *img* - Image - Image to be tracked.
* *bb* - tuple - Bounding Box tuple (x, y, w, h)
* *ts* - TrackSet - SimpleCV.Features.TrackSet.
Optional PARAMETERS:
lower - Lower HSV value for inRange thresholding
tuple of (H, S, V)
upper - Upper HSV value for inRange thresholding
tuple of (H, S, V)
mask - Mask to calculate Histogram. It's better
if you don't provide one.
num_frames - number of frames to be backtracked.
**RETURNS**
SimpleCV.Features.Tracking.CAMShift
**HOW TO USE**
>>> cam = Camera()
>>> ts = []
>>> img = cam.getImage()
>>> bb = (100, 100, 300, 300) # get BB from somewhere
>>> ts = CAMShiftTracker(img, bb, ts, lower=(40, 120, 120), upper=(80, 200, 200), num_frames=30)
>>> while (some_condition_here):
... img = cam.getImage()
... bb = ts[-1].bb
... ts = CAMShiftTracker(img, bb, ts, lower=(40, 120, 120), upper=(80, 200, 200), num_frames=30)
... ts[-1].drawBB()
... img.show()
This is too much confusing. Better use
Image.track() method.
READ MORE:
CAMShift Tracker:
Uses meanshift based CAMShift thresholding technique. Blobs and objects with
single tone or tracked very efficiently. CAMshift should be preferred if you
are trying to track faces. It is optimized to track faces.
"""
lower = np.array((0., 60., 32.))
upper = np.array((180., 255., 255.))
mask = None
num_frames = 40
if not isinstance(bb, tuple):
bb = tuple(bb)
bb = (int(bb[0]), int(bb[1]), int(bb[2]), int(bb[3]))
for key in kwargs:
if key == 'lower':
lower = np.array(tuple(kwargs[key]))
elif key == 'upper':
upper = np.array(tuple(kwargs[key]))
elif key == 'mask':
mask = kwargs[key]
mask = mask.getNumpyCv2()
elif key == 'num_frames':
num_frames = kwargs[key]
hsv = cv2.cvtColor(img.getNumpyCv2(), cv2.cv.CV_BGR2HSV)
if mask is None:
mask = cv2.inRange(hsv, lower, upper)
x0, y0, w, h = bb
x1 = x0 + w -1
y1 = y0 + h -1
hsv_roi = hsv[y0:y1, x0:x1]
mask_roi = mask[y0:y1, x0:x1]
hist = cv2.calcHist( [hsv_roi], [0], mask_roi, [16], [0, 180] )
cv2.normalize(hist, hist, 0, 255, cv2.NORM_MINMAX);
hist_flat = hist.reshape(-1)
imgs = [hsv]
if len(ts) > num_frames and num_frames > 1:
for feat in ts[-num_frames:]:
imgs.append(feat.image.toHSV().getNumpyCv2())
elif len(ts) < num_frames and num_frames > 1:
for feat in ts:
imgs.append(feat.image.toHSV().getNumpyCv2())
prob = cv2.calcBackProject(imgs, [0], hist_flat, [0, 180], 1)
prob &= mask
term_crit = ( cv2.TERM_CRITERIA_EPS | cv2.TERM_CRITERIA_COUNT, 10, 1 )
new_ellipse, track_window = cv2.CamShift(prob, bb, term_crit)
if track_window[2] == 0 or track_window[3] == 0:
track_window = bb
track = CAMShiftTrack(img, track_window, new_ellipse)
return track
