# Edge Detection in Image Processing

Topics: Image processing, Edge detection, Feature detection Pages: 2 (250 words) Published: March 19, 2013
EDGE DET ECT I ON I N I M AGE PROCESSI N G
CANNY’S EDGE DETECTOR

What are edges in an image?
Input Image Output Image

Look for places in the image where Intensity changes quickly

Edges are pixels where image brightness changes abruptly

Significance of Edge Detection

Significantly reduces the amount of data to be processed and filters out useless info, while preserving the imp. features

Steps in Canny’s Edge Detector
   

Image Smoothening Finding the gradient Non-maximum suppression Image Thresholding

Image Smoothening (blurring)
Original Image After Smoothening

•The image is convolved with gaussian filter •High frequency noise is eliminated

Finding the gradient at every pixel

Intensity value at every pixel is available :I(x,y) Calculate gradient magnitude (done by convolving the image with Sobel operator) Calculate gradient directions

Calculates gradient Gx and Gy  Total gradient: sqrt(Gx2 + Gy2)  Direction: invtan(Gy/Gx)  Gradient direction is Quantized: horizontal, vertical and diagonal 

Non- maximum suppression

  

Let M(x,y) be the gradient in a given direction M(x,y) > M(x+dx, y+dy) M(x,y) > M(x-dx, y-dy) Condition not satisfied- suppress that pixel (cannot act as an edge)

Thresholding (eliminating unwanted edges)
2 thresholds : tlow and thigh  If M(x,y) < tlow ? This pixel cannot represent an edge  If M(x,y) > thigh ? This pixel represents an edge  If tlow < M(x,y) < thigh ? check for connectivity with an edge pixel 

Thank you…
Original Image After applying Canny’s Edge detector

Please join StudyMode to read the full document