def find_hand_center(self, thresh):
"""
"""
thresh[thresh > 0] = 1
th, tw = thresh.shape
dist = dte(thresh).flatten() # find dist between hand pts and hand contour (can speed up)
distidx = dist.argmax()
return (np.mod(distidx, tw), distidx//tw)
def voronoi(self):
if self.verbose:
i = [i for i in tqdm([], desc='Voronoi Segment')]
inverted_binary_mask_stk = self.mask_stack == 0
distance_mask_stk = dte(
inverted_binary_mask_stk,
sampling=[self.pixel_size, self.pixel_size, self.z_step_size])
max_mask_stk = distance_mask_stk < self.distance_thresh
labels = watershed(image=distance_mask_stk,
markers=self.mask_stack,
mask=max_mask_stk)
self.voronoi_stack = labels
self.voronoi_images = self.stack_to_images(labels)
cY = d[cidx]['y']
cdepth = d[cidx]['d']
cinside = d[cidx]['i']
ctacking = d[cidx]['s']
cjdepth = d[cidx]['rd']
coldx = []
coldy = []
oridepth = 2000
error = 0
sr = 2 # search region
from scipy.ndimage.morphology import distance_transform_edt as dte
dismtx = np.ones((sr * 2 + 1, sr * 2 + 1))
dismtx[sr, sr] = 0
dismtx = dte(dismtx)
for i in xrange(len(X)):
print i
if i == 0:
oldx.append(X[i])
oldy.append(Y[i])
coldx.append(cX[i])
coldy.append(cY[i])
pass
else:
dist = ((X[i] - oldx[-1])**2 + (Y[i] - oldy[-1])**2)**0.5
vx = X[i] - oldx[-1]
vy = Y[i] - oldy[-1]
cvx = (cX[i] - coldx[-1]) * 1
def poolCore(inAr):
ip = imageProcessor()
imgSub = inAr[0]
mcdp = inAr[1]
mcs = inAr[2]
ct = inAr[3]
lav = inAr[4]
mcdp2 = inAr[5]
mosrr = inAr[6]
pf1 = inAr[7][0]
pf2 = inAr[7][1]
pf3 = inAr[7][2]
ppbfr = inAr[8][0]
ppbfd = inAr[8][1]
ppbd = inAr[9][0]
ppdr = inAr[9][1]
pped = inAr[10][0]
pper = inAr[10][1]
atw = inAr[11]
mcdp3 = inAr[12]
y0, y1, x0, x1, yES, xES = inAr[13][0], inAr[13][1], inAr[13][2], inAr[13][
3], inAr[13][4], inAr[13][5]
sys.stdout.write("\r" + str(ct) + " ")
sys.stdout.flush()
# print np.mean(imgSub)
imggrab = imgSub
imgSub = ip.floorBackgroundSubtraction(imgSub, minVal=lav)
openingKern = np.ones((mcdp2, mcdp2), np.uint8)
imgSub = cv2.morphologyEx(imgSub, cv2.MORPH_OPEN, openingKern)
imgSub = ip.floorBackgroundSubtraction(imgSub,
minVal=np.mean(imgSub) * mosrr)
# fig = plt.subplot2grid((2, 2), (0, 0))
# plt.imshow(imgSub,cmap='gray')
if pf1:
imgSub = cv2.bilateralFilter(np.float32(imgSub), ppbfr, ppbfd, ppbfd)
if pf2: imgSub = ip.morphBluring(imgSub, kernelSize=ppbd, repeats=ppdr)
if pf3: imgSub = ip.morphErode(imgSub, kernelSize=pped, repeats=pper)
imgSub = np.array(imgSub.astype(int), dtype=np.uint8)
imgSub = cv2.adaptiveThreshold(imgSub, 255, cv2.ADAPTIVE_THRESH_GAUSSIAN_C,
cv2.THRESH_BINARY, atw, 0)
imgSub = cv2.morphologyEx(imgSub, cv2.MORPH_CLOSE, np.ones((7, 7),
np.uint8))
# plt.subplot2grid((2, 2), (0, 1),sharex=fig,sharey=fig)
# plt.imshow(imgSub,cmap='gray')
imgSub = dte(imgSub != 0)
imgSub = dte((imgSub <= mcdp3) * (imgSub != 0))
peakList = []
for mcd in mcdp:
peakList = peakList + ip.xyLocalMaxFinder(
imgSub, minpeakSpacing=mcs, minPeakVal=mcd, negationList=peakList)
# plt.subplot2grid((2, 2), (1, 0),sharex=fig,sharey=fig)
# plt.imshow(imgSub,cmap='jet')
# plt.plot([c[1] for c in peakList],[c[0] for c in peakList],'mo',ms=4)
# plt.subplot2grid((2, 2), (1, 1),sharex=fig,sharey=fig)
# plt.imshow(imggrab,cmap='gray')
# plt.plot([c[1] for c in peakList],[c[0] for c in peakList],'ro',ms=4)
# plt.show()
# plt.pause(2.0)
# plt.clf()
y0R, y1R = y0 + yES, y1 - yES
x0R, x1R = x0 + xES, x1 - xES
peakList = [(y + y0, x + x0) for (y, x) in peakList if (
y + y0 >= y0R and y + y0 <= y1R and x + x0 >= x0R and x + x0 <= x1R)]
return peakList