Crop a specific area for the sem image

I have many images containing SEMs and annotations. I want to crop out just the SEM content, the black square regions, using Matlab. This requires Matlab to locate the black borders of the SEMs and record those coordinates, then use imcrop to crop the specific areas. I don't want the white content within the SEMs to affect the determination.

thanks!

[KSSV]

Why don't you show us what areas you want from the image. You may highlite the regions using paint.

[luo jingcheng]

the whole black section with sclerites in it

Answer 1

[Akira Agata]

How about the following?

% Read the image
I = imread("https://www.mathworks.com/matlabcentral/answers/uploaded_files/1522566/image.png");

% Create ROI mask
Igray = rgb2gray(I);
BW = Igray>50;
BW = imclearborder(BW);
BW = bwareafilt(BW, [5000 Inf]); % Ignore small ROIs (<5000 pixel)

% Calculate bounding box for each ROI
s = regionprops("table", BW);
s = sortrows(s, "Area", "descend");

% Crop each region from the image
cImg = cell(height(s), 1);
for kk = 1:height(s)
  cImg{kk} = imcrop(I, s.BoundingBox(kk, :));
end

% Visualize the result
figure
montage(cImg, "BackgroundColor", "w")

[luo jingcheng]

that is sooo helpful!

In fact, I initially planned to separate the entire SEM image before croping each of the image because I was concerned that the text below might affect the results during the binarization process. you have directly implemented this step. Now, I intend to adjust the background of each image to be completely black (some image contain adjacent sclerites)

Answer 2

[Image Analyst]

See my Image Segmentation Tutorial in my File Exchange:

Image Analyst's File Exchange

It's a generic, general purpose demo of how to threshold an image to find blobs, and then measure things about the blobs, and extract certain blobs based on their areas or diameters. Note how each of the coins in the demo image is cropped out to its own image. Just replace the image with yours. And you can use bwareafilt to get rid of small blobs, like the letters or other annotations in the image.