def startFrame():
# Read a frame from feed
global frame
ret, frame = cap.read()
# Convert to greyscale frame
global frame_gray
frame_gray = cv2.cvtColor(frame, cv2.COLOR_BGR2GRAY)
# Get histogram of frame
hist_img = cv2.calcHist([frame], [0], None, [256], [0, 256])
#plt.plot(hist_img)
#plt.show()
# Pass frame to histogram adjustment to remove ouliers
hist_no_outliers, lower_index = outliers.removeOutliersThresh(hist_img)
#plt.plot(hist_no_outliers)
#plt.show()
# Pass histogram to adaptive thresholding to determine level
global threshLevel
threshLevel = thresh.bi_level_img_threshold(hist_no_outliers)
# Adjust start index of hist and add manual level adjustment
global threshLevelAdjust
threshLevelAdjust = threshLevel + lower_index + 15
#print('Bi level thresh', threshLevelAdjust)
struct_el = cv2.getStructuringElement(cv2.MORPH_ELLIPSE,(25,25))
global frame_open
frame_open = cv2.morphologyEx(frame_gray, cv2.MORPH_OPEN, struct_el)
#cv2.imshow('open',frame_open)
return frame_open
def imgThreshold(frame):
# Convert to greyscale frame
frame_gray = cv2.cvtColor(frame, cv2.COLOR_BGR2GRAY)
#cv2.imshow('original',frame_gray)
# Create structuring element - disk to remove glint
struct_el = cv2.getStructuringElement(cv2.MORPH_ELLIPSE,(35,35))
frame_open = cv2.morphologyEx(frame_gray, cv2.MORPH_OPEN, struct_el)
#cv2.imshow('open',frame_open)
# Get histogram of frame
# more efficient than calcHist and eliminates memory error
hist_img = np.bincount(frame_gray.flatten())
#print('len_hist',len(hist_img))
# truncate histogram to remove bin 0 and bin 255 instead of removing outliers which
# removes all the bins
hist_img = hist_img[1:len(hist_img)-2]
# Pass histogram to adaptive thresholding to determine level
threshLevel = thresh.bi_level_img_threshold(hist_img)
#if(threshLevel > 100):
threshLevel = 30
# Adjust start index of hist and add manual level adjustment
#threshLevelAdjust = threshLevel + lower_index
#print('Bi level thresh', threshLevelAdjust)
# Threshold frame using level obtained from adaptive threshold
#print ("thresh level %d " % threshLevel)
ret,threshPupil = cv2.threshold(frame_open,threshLevel,255,cv2.THRESH_BINARY)
#cv2.imshow('thresh pupil',threshPupil)
# Invert and threshold frame to isolate only glint
#frameInv = np.invert(frame_gray)
#cv2.imshow('frameInv',frameInv)
ret,threshGlint = cv2.threshold(frame_gray,200,255,cv2.THRESH_BINARY_INV)
#cv2.imshow('thresh glint',threshGlint)
return threshPupil, threshGlint
# Convert to greyscale frame
frame_gray = cv2.cvtColor(frame, cv2.COLOR_BGR2GRAY)
i = i+1
print('i ', i)
# Get histogram of frame
hist_img = cv2.calcHist([frame_gray], [0], None, [256], [0, 256])
#plt.plot(hist_img)
#plt.show()
# Pass frame to histogram adjustment to remove ouliers
hist_no_outliers, lower_index = outliers.removeOutliersThresh(hist_img)
#plt.plot(hist_no_outliers)
#plt.show()
# Pass histogram to adaptive thresholding to determine level
threshLevel = thresh.bi_level_img_threshold(hist_no_outliers)
# Adjust start index of hist and add manual level adjustment
threshLevelAdjust = 40 #threshLevel + lower_index
print('Bi level thresh', threshLevelAdjust)
plt.plot(hist_no_outliers)
plt.show()
struct_el = cv2.getStructuringElement(cv2.MORPH_ELLIPSE,(25,25))
frame_open = cv2.morphologyEx(frame_gray, cv2.MORPH_OPEN, struct_el)
cv2.imshow('open',frame_open)
# Threshold frame using level obtained from adaptive threshold
ret,frameBinary = cv2.threshold(frame_open,threshLevelAdjust,255,cv2.THRESH_BINARY)
cv2.imshow('binaryOrig',frameBinary)
