Applying a small kernel to an image to reduce blurriness.

unblur.m

% Demo of a basic filter for deblurring an image

% Grab image
img_url = 'http://www.artwallpaperhi.com/thumbnails/detail/20121017/sand%20macro%20blurred%201920x1200%20wallpaper_www.artwallpaperhi.com_77.jpg';
img = imread(img_url);
img = single(img) / 255;

% Apply convolution. Play around with the filter matrix if you're curious.
filter = [-1 -1 -1; -1 9 -1; -1 -1 -1];
result = zeros(size(img));
padded = padarray(img, [1 1 0], 'replicate');
for c=1:size(img, 3)
    convolved = conv2(padded(:, :, c), filter);
    result(:, :, c) = convolved(3:end-2, 3:end-2);
end

% Show everything
subplot(1, 2, 1);
imshow(img);
text(0, -12, 'Before');
axis equal;
subplot(1, 2, 2);
imshow(result);
text(0, -12, 'After');
axis equal;