Cut an area in the image

Hey everyone;

In fact my problem is I have an image that I want to automatically remove black areas as shown in attachment. http://www.casimages.com/img.php?i=120531055451175771.png What I meant exactly what is extracting the part surrounded by yellow, so I can do processing on that image.

I thought imcrop, but in this case, the selection must be done manually.

thank you

Answer 1

[Andrei Bobrov]

use function imcrop from Image Processing Toolbox

Answer 2

[Elad]

you can choose top left and bottom left boundaries using range=ginput(2); and then image=image(range(1,2):range(2,2),range(1,1):range(2,1),:);

Answer 3

[Image Analyst]

What I would do, assuming your corners of your sheared image touch the borders of the image are to scan each edge of the image looking for non-zero pixels. Then sort the row values and sort the column values of the corner locations. Then you want to go between the 2nd and 3rd values.

row1 = sortedRows(2);
row2 = sortedRows(3);
col1 = sortedCols(2);
col2 = sortedCols(3);
% Now crop
height = row2-row1+1;
width = col2-col1+1;
croppedImage = imcrop(fullImage, [col1, row1, width, height]);

[nomad nomad]

The image is the output of a program that the projective transformation of an input image and since I have no prior knowledge of the shape of the image, I can not define the coordinates.

[Image Analyst]

So what? You don't need to. Just scan the borders, like

leftColumn = fullImage(:, 1); % Get column 1

% Find where it's non-zero:

upperLeftRow = find(leftColumn >0, 1, 'first');

Similar for the other 4 corners that touch the border. Where in that are you required to have prior knowledge of any transform that was used to warp the image?