Edge Detection with Blur using Python OpenCV

Edge detection identifies object boundaries by detecting sharp intensity changes in an image. When applied directly to real-time frames, it often produces noisy and overly sharp edges. To improve results, blurring is applied as a preprocessing step before edge detection. Blurring before edge detection helps in:

• Reduces noise
• Smoothens intensity variations
• Produces cleaner and more meaningful edges

Hence, blurring is an important step in most real-world edge detection applications.

Functions and Methods Used

1. cv2.flip(): This function flips an image or video frame, which is used to create a mirrored view. Below is the syntax:

cv2.flip(src, flipCode)

Parameters:

• src: Input image or frame
• flipCode: 1 for horizontal flip

2. cv2.blur(): This function applies an averaging blur to an image. Below is the syntax:

cv2.blur(src, ksize)

Parameters:

• src: Input image
• ksize: Kernel size as (width, height)

3. cv2.Canny(): This function detects edges in an image using Canny edge detection algorithm. It identifies areas with strong intensity changes that represent object boundaries. Below is the syntax:

cv2.Canny(image, threshold1, threshold2)

Parameters:

• image: Input image
• threshold1: Lower threshold value
• threshold2: Upper threshold value

Python Implementation

The following program captures live video from a webcam, applies blur and edge detection and displays different stages of processing in real time for comparison.

• cv.VideoCapture(0) opens the default webcam.
• cam.read() continuously captures frames from the camera.
• cv.flip(frame, 1) mirrors the frame for natural viewing.
• cv.blur(frame, (3,3)) smoothens the image and removes noise.
• cv.Canny() performs edge detection on both original and blurred frames.
• Multiple cv.imshow() windows display all processing stages.
• Pressing Escape (27) exits the program and releases resources.

import cv2 as cv

cam = cv.VideoCapture(0)

while True:
    ret, frame = cam.read()
    frame = cv.flip(frame, 1)

    blur = cv.blur(frame, (3, 3))

    edges_original = cv.Canny(frame, 100, 200)
    edges_blur = cv.Canny(blur, 100, 200)

    cv.imshow("Original Frame", frame)
    cv.imshow("Blurred Frame", blur)
    cv.imshow("Edge Detection on Original", edges_original)
    cv.imshow("Edge Detection on Blurred", edges_blur)

    key = cv.waitKey(1)
    if key == 27:  # Escape key
        break

cam.release()
cv.destroyAllWindows()

Output: After running the program, four live windows are displayed:

• Original webcam frame
• Blurred frame
• Edge detection without blur
• Edge detection with blur

The blurred edge detection output clearly shows reduced noise and smoother edges.