OpenCV - Adaptive-thresholding / effective noise reduction?

Question

I am new to OpenCV and I just read about cv2.adaptiveThreshold() and decided to try it out. Sadly there is this huge amount of noise I cannot seem to get rid of.

What are some effective ways to reduce noise so I can draw proper contours? What is the best practice and why?

Here is the snippet:

 import cv2 import numpy as np ##################################### winWidth = 640 winHeight = 840 brightness = 100 cap = cv2.VideoCapture(0) cap.set(3, winWidth) cap.set(4, winHeight) cap.set(10, brightness) kernel = (5, 5) ########################################### def preprocessing(frame): imgGray = cv2.cvtColor(frame, cv2.COLOR_BGR2GRAY) # mask = cv2.inRange(imgHsv, lower, upper) imgBlurred = cv2.GaussianBlur(imgGray, kernel, 1) gaussC = cv2.adaptiveThreshold(imgBlurred, 255, cv2.ADAPTIVE_THRESH_GAUSSIAN_C, cv2.THRESH_BINARY_INV, 11, 2) imgDial = cv2.dilate(gaussC, kernel, iterations=3) imgErode = cv2.erode(imgDial, kernel, iterations=1) return imgDial def getcontours(imPrePro): contours, hierarchy = cv2.findContours(imPrePro, cv2.RETR_TREE, cv2.CHAIN_APPROX_NONE) for cnt in contours: cv2.drawContours(imgCon, cnt, -1, (255, 0, 0), 3) ################################################### while (cap.isOpened()): success, frame = cap.read() if success == True: frame = cv2.flip(frame, 1) imgCon = frame.copy() imPrePro = preprocessing(frame) getcontours(imPrePro) cv2.imshow("Preprocessed", imPrePro) cv2.imshow("Original", imgCon) if cv2.waitKey(1) & 0xFF == ord("q"): break

Answer 1

I think it is best if we look at the blockSize and C parameters.

Form the source :

blockSize: Size of a pixel neighborhood that is used to calculate a threshold value for the pixel: 3, 5, 7, and so on.

C: Constant subtracted from the mean or weighted mean (see the details below). Normally, it is positive but may be zero or negative as well.

In your example you set C to 2:

gaussC = cv2.adaptiveThreshold(imgBlurred, 255, cv2.ADAPTIVE_THRESH_GAUSSIAN_C, cv2.THRESH_BINARY_INV, 11, 2)

C=2	C=8

As you can see, you need to play with blockSize and C parameters to get the desired result from adaptive-threshold.

In this question, we achieve less-noise by increasing C parameter.