Error in my edge detection function?

I'm attempting to create a canny edge detector and I've run into a speed bump while checking for the edge pixels. Below I have the code I'm using to check which pixel is on an edge but I keep getting the error "Attempted to access mag(0,18); index must be a positive integer or logical" Is there a way to fix this? For this code, mag=gradient magnitude, dir=gradient direction, img=image, and edgecand=the edge candidate.

edgecand = mag;
for x=1:size(img,1);
    for y=1:size(img,2);
        if (dir(x,y)==1);
            if(mag(x,y)>mag(x-1,y))&&(mag(x,y)>mag(x+1,y));
                %happy
            else edgecand(x,y)=0;
            end
        end
        if (dir(x,y)==2);
            if(mag(x,y)<mag(x+1,y+1))&&(mag(x,y)>mag(x-1,y-1));
                %happy
            else edgecand(x,y)=0;
            end
        end
        if (dir(x,y)==3);
            if(mag(x,y)>mag(x,y-1))&&(mag(x,y)<mag(x,y+1));
                %happy
            else edgecand(x,y)=0;
            end
        end
        if (dir(x,y)==4);
            if(mag(x,y)>mag(x+1,y-1))&&(mag(x,y)<mag(x-1,y+1));
                %happy
            else edgecand(x,y)=0;
            end
        end
    end
end

Thank you for any help!

[Abigail]

I forgot to mention that the dir(x,y) 1, 2, 3, and 4 stand for the horizontal,vertical etc. directions.

Answer 1

[Image Analyst]

You can't have 0 as an index. There is no 0th row. MATLAB is 1-based array indexing, not 0-based like C.

[Abigail]

Oh,I see. Thank you!

Answer 2

[Jyoti Verma]

c=imread('carplate1.jpg');

b=rgb2gray(c);

px=[-1 0 1;-1 0 1;-1 0 1]; %_horizontal edges,

b1=filter2(px,b);

figure,imshow(b1/255)

py=px'; %%_vertical edges,

b2=filter2(py,b);

figure,imshow(b2/255)

edge_p=sqrt(b1.^2+b2.^2);

figure,imshow(edge_p/255)

%figure,imshow(edge_p/512), Dividing by 512 darkens the image

%figure,imshow(edge_p/128), Dividing by 128 Lighten the image,

edge_t=im2bw(edge_p/255,0.3); %a binary image containing edges only can be produced by thresholding.

figure,imshow(edge_t)

%edge_g=im2bw(edge_p/255,0.3);

%figure,imshow(edge_g)

%We can obtain edges by the Prewitt _filters

edge_k=edge(b,'prewitt');

figure,imshow(edge_k)

%We can obtain edges by the Prewitt _filters

edge_s=edge(b,'sobel');

figure,imshow(edge_s)