Binarizing image with single pixels

I have some photos of random pixels marked on a low noisy background. I want to binarize the image, in order to have a clear array of 0s and 1s, and I would like to make each spot as one element of the matrix.

What I did was to grayscale the image, and then try to binarize it. Unfortunately, it is not enough. Adding Adaptive histeresys equalisation (Ae = adapthisteq(Ag);) makes it worst. Maybe strel function could help?

Thanks

This is my code. I think code will be similar to datamatrix and QR code reading.

Thank you

if dim ~= 1

Ag = rgb2gray(photo);

else

Ag = photo;

end

Ab=imbinarize(Ag);

Answer 1

[Animesh]

Hi @Mat

If you're working with an image that has random pixels against a faintly noisy background and wish to detect those random pixels, you can employ morphological operations to enhance the features of the spots and distinguish them from the background noise.

A suggested approach you can try is:

Step 1: Convert the image to greyscale (if not already).

Step 2: Apply a filter to reduce noise.

Step 3: Binarize the image using thresholding.

Step 4: Use morphological operations to clean up the image and isolate the spots.

You can refer to the following documentation on morphological operations:

• Morphological operations: https://in.mathworks.com/help/images/morphological-filtering.html
• Structuring Elements: https://in.mathworks.com/help/images/structuring-elements.html