HARRIS_CORNER_CVIP

harris_corner_cvip() - a spatial-domain method for corner detection.

SYNTAX
DESCRIPTION
REFERENCE
EXAMPLE
CREDITS

SYNTAX

[ CRF ] = harris_corner_cvip( input_img , threshold, std, alfa, maxsupress)

Input Parameters include:

input_img - The input image. Can be multiband

threshold - A rational value in the interval [0,1]. It is used in the thresholding step as a coefficeient multiplied to the maximum gray level in the input image.

std - The Standard Deviation of the Gaussian blur kernel. the size of the gaussian is related to std by: size = 4*std Thus we have 2 standard deviation of the gaussian inside the kernel.

alfa - Sensitivity value.

maxsupress - Option to supress non-maxima either in a neighbourhood or based on direction. It is either '1' or '2'. Default: 2.

DESCRIPTION

This is developed by Harris ans Stephens in 1998.This method consists of four steps 1. Blur the image with 2-D Gaussian convolution mask. 2. Find the approximate brightness gradient in two perpendicular directions. 3. Blur two brightness results with 2-D Gaussian. 4. Find the corner response function(CRF). 5. Threshold the CRF and apply non maxima supression.

REFERENCE

1.Scott E Umbaugh. DIGITAL IMAGE PROCESSING AND ANALYSIS: Applications with MATLAB and CVIPtools, 3rd Edition.

EXAMPLE

% original image

 I = imread('Shapes.bmp');

% threshold value from [0,1]

 threshold = .01;

% standard deviation

 std = 4;

% alfa

 alfa = 0.5;

%Option to supress non-maxima either in a neighbourhood  or based on direction.

 maxsupress = 2;

% resultant Corner response

 [ CRF ] = harris_corner_cvip( I , threshold, std,alfa, maxsupress);

% Display input image

 figure; imshow(I);title('Input image')

% Display output image

 figure; imshow(CRF);title('Resultant image from corner response function');

The size of the gaussian is: 16 
The threshold value is: 109468.843153

HARRIS_CORNER_CVIP

Contents

SYNTAX

DESCRIPTION

REFERENCE

EXAMPLE

CREDITS