Basic Method to Detect Edges of an Image

As you may have studied in subjects like Operation Research or might have experienced, the changes occur only at edges. And hence to detect these edges we need to track the drastic change in pixel value. Boundary or an edge will have a sudden abrupt change in pixel value.

In the first stage of edge detection, the image is cleaned from noise by applying some filters. But also need to remember that not all edges will have a distinct change from its neighboring pixels. So there are problems with false edge detection, missing actual edges etc.

Two major methods widely employed in Edge Detection  :-


  • Derivative Method :-  In this method, we take the derivative w.r.t to time. We know that derivative yields maximum location which might correspond to an edge. If we take the same derivative one more time i.e twice on original, maximum turns to zeros. Hence one more approach it to take second derivative and look for maximum zeros. This is nothing but Laplacian Approach.
  • Gradient Method :- Here it looks at nearby pixels to find the changes in pixel value. If value corresponding to them are large, then it is a probable edge.

Comments

Post a Comment

Popular posts from this blog

Mini Project for Image Processing Beginners

Program to convert image byte array to PDF

Program to convert base64 to Image and vice versa