MATLAB: Have I done something wrong with this code for edge detection on extreme gradient value

edge detectionImage Processing Toolbox

I have first translated from RGB to HSV and separated to hue and saturation and intensity component h(i,j),s(i,j),v(i,j)
5*5 boundary detection operator window with original image f(i,j) and the central pixel of the window (i,j) is the boundary pixel to be detected,(x,y) is one of the eight neighborhood pixels of the central pixel
the central pixel's hue saturation and brightness in the direction of theta(0 to 315 with 45 degree spacing each) are defined as:
hθ(i, j)= h(x, y)-h(i, j)
sθ(i, j)= s(x, y-s (i, j)
vθ(i, j)=v(x, y)-v(i, j)
   w=[1,1,1,1,1;1,1,1,1,1;1,1,-24,1,1;1,1,1,1,1;1,1,1,1,1];
gradient is
gθ(i, j)= hθ(x, y)+ sθ(i, j)+ vθ(x, y);
code
function g=exgradient(hsv)
hsv=rgb2hsv(rgb);
h=hsv(:,:,1);
s=hsv(:,:,2);
v=hsv(:,:,3);
theta=1:45:315
w=[1,1,1,1,1;1,1,1,1,1;1,1,-24,1,1;1,1,1,1,1;1,1,1,1,1]; 
for i=1:183
      for j=1:276
          h(theta)*(i,j)=h(x,y)-h(i,j);
          s(theta)*(i,j)=s(x,y)-s(i,j);
          v(theta)*(i,j)=v(x,y)-v(i.j);
      end
end

Best Answer

Looks like 3x3 to me, not 5x5. Just have your kernel be
kernel = [-1 -1 -1;-1 8 -1;-1 -1 -1] / 9; % Average pixel difference.
edgeImage = conv2(double(grayImage), kernel, 'same'); % the filter.
Related Question