Closing Operator

Building the Cantata Workspace


The closing is an elementary operator that accepts as parameters a function and a structural element. This operator applies for signals, binary and gray scale images.

The structural elements can be criated by the following ways:

  1. using the tool Struc. El. 3x3 to generate 3x3 planar elements;

  2. using the tool Disk str. el. to generate any disk, using three distance metrics: euclidean, city-block and chess-board.

  3. Using the tool Viff-> Str.El. to convert any viff file to a structural element.


To build and experiment a workspace that performs some openings:

  1. select the $MMACH/viff/blobs2.xv image;

  2. apply the negation operation to the input image;

  3. visualize this image using display image;

  4. create a family of four structural elements of increasing size, using the vdiskstr tool with the option Euclidian distance;

  5. perform the closing of the negation of the input image by these disks;

  6. convert the negation of the input image and its four closings to the BYTE type;

  7. add all the five images;

  8. display the result of the addition using a random color map.


Exercises

  1. Apply the closing operator on the 1D signal $MMACH/viff/signal1.viff. Use flat and non flat structural elements. What differences could you note between the efect of the flats and non flats structural elements?

  2. Apply the closing by the 3x3 square to the image $MMACH/viff/letters_noisy.xv.

  3. Apply to the image $MMACH/viff/letters_noisy.xv a family of four closings. Chouse as structural elements the four segments of three points including in the 3x3 square and with direction o, 45, 90 and 135 degrees. Perform the intersection of the result of the four closings.

  4. Compare the results of the two previous exercices. What are the differeces between the two filters?

  5. Apply an closing by an octagon (compostion of the cross 3x3 and the square 3x3) to the negation of the image $MMACH/viff/disk_square.xv and take the difference with the original image.

  6. Could you see an application for the previous exercise?