Scribd
Upload
2 views13 pages

Image Processing (DSP Research)

The document discusses various edge detection techniques in image processing, including Sobel, Prewitt, Laplacian, and Canny methods. Each technique is explained with its mathematical foundations, implementation in Python, and specific use cases. The document also includes references for further reading on edge detection applications.

Uploaded by

Lakshay Dhahinwal
Copyright
© © All Rights Reserved
We take content rights seriously. If you suspect this is your content, claim it here.
Available Formats
Download as PPTX, PDF, TXT or read online on Scribd
2 views13 pages

Image Processing (DSP Research)

The document discusses various edge detection techniques in image processing, including Sobel, Prewitt, Laplacian, and Canny methods. Each technique is explained with its mathematical foundations, implementation in Python, and specific use cases. The document also includes references for further reading on edge detection applications.

Uploaded by

Lakshay Dhahinwal
Copyright
© © All Rights Reserved
We take content rights seriously. If you suspect this is your content, claim it here.
Available Formats
Download as PPTX, PDF, TXT or read online on Scribd
You are on page 1/ 13

INDIAN INSTITUTE OF TECHNOLOGY ROORKEE

Edge Detection in Image Processing


ECN-511

Lakshay Dhahinwal
(Enrollment No. 24531004)
What is an Edge?

• Edge represents a significant change in intensity or colour


between adjacent pixels over a short distance.

2
Edge Detection Techniques

• Sobel Edge Detection -


– Calculates gradient in two directions: horizontal(Gx) and vertical(Gy).
– Convolution kernels are used to compute these gradients.
– Gx = . Gy =

– The image is made into a matrix using NumPy module in Python, then
convolution is done with these filters to obtain 2 Kernels.

– Magnitude of the gradient is calculated is calculated at each pixel.

3
Sobel Edge Detection

– Orientation of the edge is calculated by the formula:

=)
– Finally, a threshold is applied to get the final edge map.

• These steps can be demonstrated in a Python Script.

4
Sobel Edge Detection

Sobel Edge Detection

5
Edge Detection Techniques

• Prewitt Edge Detector -


– Similar to Sobel, but here the weighing of the pixels is not done, which can
lead to varying results depending on image selection.
– Prewitt Kernels:
• Gx = . Gy =

– Used when simpler, noise free images are present. Less computationally
demanding.
– Python Implementation:

6
Prewitt Edge Detection

7
Edge Detection Techniques

• Laplacian Based Edge Detection –


– It is a double derivative-based edge detection.
– A gaussian blur is applied to the image to reduce noise.
– The Laplacian kernel represents the Laplacian Operation
/
– Laplacian Kernel:

– This kernel is multiplied by the image matrix.


– Python Implementation:

8
Laplacian Based Edge Detection

9
Edge Detection Techniques

• Canny Edge Detection-


– Improvement on Sobel Edge Detection by removing all the edges that are
not required. (Output of Sobel = Input of Canny)
– Thins the edges to about 1 pixel wide.
– Apply Hysteresis Thresholding.
• 2 Level Thresholding.
• 2 threshold values taken, if the gradient magnitude is higher than high
threshold they are marked as strong edges, between the two thresholds it’s a
weak edge and below the thresholds it is discarded as being noise.
– Edge Tracking – weak edges only considered if they are connected string
edges.
– Python Implementation:

10
Canny Edge Detection

11
References
• P. Ganesan and G. Sajiv, "
A comprehensive study of edge detection for image processing
applications
," 2017 International Conference on Innovations in Information,
Embedded and Communication Systems (ICIIECS), Coimbatore, India,
2017, pp. 1-6, doi: 10.1109/ICIIECS.2017.8275968.
• Finding the Edges (Sobel Operator) – Computerphile
• Canny Edge Detector - Computerphile
• Edge Detection Using Laplacian | Edge Detection

12
Thank You

13

You might also like