MATLAB ERROR for logic operation in strel?

I have a code functions for dilation and erosion without using built-ins. When I run the code in both dilation and erosion function parts the logic operation for strel gave an error. How to solve it?

Errors I get on these lines:

dilated_val = any(struct_et(:) | sub_matrix(:));

eroded_val = all(struct_et(:) == sub_matrix(:));

The entire code is given below:

% Read the car plate image
carplate = imread('Figure1.png');
carplate_bw = imbinarize(rgb2gray(carplate));
struct_et = strel('disk', 5);

% Perform dilation operation
dilated_img = dilation(carplate_bw, struct_et);

% Perform erosion operation
eroded_img = erosion(dilated_img, struct_et);

% Display the original image, dilated image, and eroded image
figure;
subplot(1,3,1); imshow(carplate); title('Original Image');
subplot(1,3,2); imshow(dilated_img); title('Dilated Image');
subplot(1,3,3); imshow(eroded_img); title('Eroded Image');

function dilated_img = dilation(src_img, struct_et)
[src_h, src_w] = size(src_img);
[et_h, et_w] = size(struct_et);
src_img_padded = padarray(src_img, [et_h-1, et_w-1], 'replicate', 'both');
dilated_img = zeros(src_h, src_w);

for i = 1:src_h
    for j = 1:src_w
        if src_img(i,j) == 1
            % Get the sub-matrix corresponding to the structuring element
            sub_matrix = src_img_padded(i:i+et_h-1, j:j+et_w-1);
            % Perform OR operation between the sub-matrix and the structuring element
            dilated_val = any(struct_et(:) | sub_matrix(:));
            % Set the output pixel to 1 if any of the corresponding pixels in the sub-matrix is 1
            dilated_img(i, j) = dilated_val;
        end
    end
end
end

function eroded_img = erosion(src_img, struct_et)
[src_h, src_w] = size(src_img);
[et_h, et_w] = size(struct_et);
src_img_padded = padarray(src_img, [et_h-1, et_w-1], 'replicate', 'both');
eroded_img = ones(src_h, src_w);

for i = 1:src_h
    for j = 1:src_w
        % Get the sub-matrix corresponding to the structuring element
        sub_matrix = src_img_padded(i:i+et_h-1, j:j+et_w-1);
        % Perform AND operation between the sub-matrix and the structuring element
        eroded_val = all(struct_et(:) == sub_matrix(:));
        % Set the output pixel to 0 if any of the corresponding pixels in the sub-matrix is 0
        eroded_img(i,j) = eroded_val;
    end
end
end

Answer 1

[Walter Roberson]

struct_et = strel('disk', 5);

That is 1 x 1 object of class strel

dilated_img = dilation(carplate_bw, struct_et);

That is passed as the second parameter to dilation()

function dilated_img = dilation(src_img, struct_et)

which receives it under the name struct_et

            dilated_val = any(struct_et(:) | sub_matrix(:));
 

With it being a 1 x 1 object of strel class, struct_et(:) is the same as struct_et without the (:)

You are then trying to use the logical or operator | with an object of class strel but logical or is not defined for that class.

Perhaps

            dilated_val = any(struct_et.Neighborhood(:) | sub_matrix(:));

[Yigit Goktas]

dilated_val = any(struct_et.Neighborhood(:) | sub_matrix(:));

That fixed the problem for dilated_val

That fixing gave a nice dilated image out thanks for help but

I tried to apply the same thing for

eroded_val = all(struct_et(:) == sub_matrix(:));

wrote it as eroded_val = all(struct_et.Neighborhood(:)(:) == sub_matrix(:));

on the second function. However the resulting erosed image is fully black only

[Yigit Goktas]

Correction, wrote it as:

eroded_val = all(struct_et.Neighborhood(:) == sub_matrix(:));

[Walter Roberson]

Erosion should only pay attention to the pixels selected by the mask.

If you were to use

mask = logical(struct_et.Neighborhood);

then

all(sub_matrix(mask))