How could I count the number of the light small clusters in a feature map

Question

Suppose I have a heat map like this:

The value of the pixels on the heat map ranges from 0-1. The closer the value to 0 the darker the pixel is. There are six write clusters on it, as shown in the picture. Is there any method for me to know how many light clusters are there in the heatmap?

Answer 1

Source => Detection

---

My method is: Gray => Threshold => FindContours (filter by area if necessary)

#!/usr/bin/python3
# 2019/03/01
import cv2

img = cv2.imread("featmap.png")

# Gray => Threshold => FindContours (filter by area if necessary)
gray = cv2.cvtColor(img, cv2.COLOR_BGR2GRAY)
blured = cv2.medianBlur(gray, 3)
th, threshed = cv2.threshold(blured, 100, 255, cv2.THRESH_OTSU|cv2.THRESH_BINARY)
cnts = cv2.findContours(threshed, cv2.RETR_EXTERNAL, cv2.CHAIN_APPROX_NONE)[-2]

print("nums: {}".format(len(cnts)))

# draw on the original
cv2.drawContours(img, cnts, -1, (0, 255, 0), 1, cv2.LINE_AA)
cv2.imwrite("dst.png", img)

Be carefully when using findContours: How to use `cv2.findContours` in different OpenCV versions?

Answer 2

You can apply cvThreshold operation to make B/W image.

Then filter small defects with opening morhological operation

Now find connected components and get number of them.