counting cells after watershed segmentation -- openCV/Python

Question

I followed this tutorial given on watershed segmentation to separate the brown cells on the attached image. It went well (cells are separated by blue boundary) but now I would like to count those cells and determine their sizes(pixel number) in order to plot a distribution function. Could you please help how to do it?

Code is given below.

import numpy as np
import cv2

img = cv2.imread('test watershed.tif')
gray = cv2.cvtColor(img,cv2.COLOR_BGR2GRAY)
ret, thresh = cv2.threshold(gray,0,255,cv2.THRESH_BINARY_INV+cv2.THRESH_OTSU)

# noise removal
kernel = np.ones((3,3),np.uint8)
opening = cv2.morphologyEx(thresh,cv2.MORPH_OPEN,kernel, iterations = 2)

# sure background area
sure_bg = cv2.dilate(opening,kernel,iterations=3)

# Finding sure foreground area
dist_transform = cv2.distanceTransform(opening,cv2.DIST_L2,5)
ret, sure_fg = cv2.threshold(dist_transform,0.1*dist_transform.max(),255,0)


# Finding unknown region
sure_fg = np.uint8(sure_fg)
unknown = cv2.subtract(sure_bg,sure_fg)

# Marker labelling
ret, markers = cv2.connectedComponents(sure_fg)

# Add one to all labels so that sure background is not 0, but 1
markers = markers+1

# Now, mark the region of unknown with zero
markers[unknown==255] = 0

markers = cv2.watershed(img,markers)
img[markers == -1] = [255,0,0]

**UPDATE**

#thresholding a color image, here keeping only the blue in the image
th=cv2.inRange(img,(255,0,0),(255,0,0)).astype(np.uint8)


#inverting the image so components become 255 seperated by 0 borders.
th=cv2.bitwise_not(th)

#calling connectedComponentswithStats to get the size of each component
nb_comp,output,sizes,centroids=cv2.connectedComponentsWithStats(th,connectivity=4)

#taking away the background
nb_comp-=1; sizes=sizes[0:,-1]; centroids=centroids[1:,:]

bins = list(range(np.amax(sizes)))

#plot distribution of your cell sizes.

numbers = sorted(sizes)


plt.hist(sizes,numbers)

cv2.imwrite("test watershed result",img)

Answer 1

You did the hard part ! Now just threshold your result (colorwise) and call the handy connectedComponentsWithStats

#thresholding a color image, here keeping only the blue in the image
th=cv2.inRange(img,(255,0,0),(255,0,0)).astype(np.uint8)

#inverting the image so components become 255 seperated by 0 borders.
th=cv2.bitwise_not(th)

#calling connectedComponentswithStats to get the size of each component
nb_comp,output,sizes,centroids=cv2.connectedComponentsWithStats(th,connectivity=4)

#taking away the background
nb_comp-=1; sizes=sizes[1:,-1]; centroids=centroids[1:,:]

#plot distribution of your cell sizes (using matplotlib.pyplot as plt)
plt.hist(sizes)