plaplacian

Computes the Laplacian magnitude.

Synopsis

plaplacian connexity [-m mask] [im_in|-] [im_out|-]

Description

plaplacian computes the Laplacian of an image. The Laplacian is an approximation of the second derivative of an image.

This operator can be used to detect closed contours in an image. But it is very sensitive to noise.

The algorithm uses the convolution with the following kernel:

for 4-connexity

   |+0 -1  +0|
   |-1  +4 -1|
   |+0 -1  +0|

for 8-connexity

   |-1 -1- 1|
   |-1  +8 -1|
   |-1 -1 -1|

The output image im_out is of the same type as the input image im_in.

For Uchar images, the values are shifted by 127 so the 0 is 127 and value < 0 is supposed negative and values > 0 is supposed to be positive. The zero crossing detection needs to use operator pzerocross with parameter value 127.

For Signed Long images, the values are not shifted. The zero crossing detection needs to use operator pzerocross with parameter value 0.

Parameters

connexity specifies the connexity neighbor relation {4 or 8}.

Inputs

im_in: a 2D grayscale image.

Outputs

im_out: an 2D grayscale image.

Result

Returns SUCCESS or FAILURE.

Examples

Edge detection using LOG algorithm (Laplacian Of Gaussian)

   pgaussfilter 1.2 tangram.pan a.pan
   plaplacian 8  a.pan b.pan
   pzerocross 8 127 b.pan out.pan

C++ prototype

Errc PLaplacian( const Img2duc &im_in, Img2duc &im_out, int connexity );

Version française

Approximation du Laplacian d'une image par convolution.

Author: Régis Clouard