MATLAB: Recognizing shapes in (binary) matrix

image processingImage Processing Toolbox

Hey all,

I'm trying to find a shape in a binary image.

In order to do that i am using bwboundaries() function to identify all the shapes in the image. I get a cell with matrices (the size of X by 2), of all the objects. Each matrix represent the number of pixels in the object.

For example, lets say there are two matrices, almost the same size. One representing the shape of a line, the other the shape of a circle (not exactly a line or a circle).

My question is – can i differ the line from the circle ?

Any help appreciated.

Yair

Best Answer

If you KNOW that you are either dealing with a line or a circle, then could you look at the MajorAxisLength and MinorAxisLength from RegionProps. Their ratios should be wildly different, yes?

Doug

Related Solutions

MATLAB: Create mask that separates objects from white background

Hi, maybe the code below can help start you off! It's not a perfect segmentation, and I'm fairly new to MATLAB so someone else might have a better method. This link may also help you too, have a look under "Functions", then "Segmentation Techniques".

apple = imread("apple.jpg"); % read in image name
grayapple = rgb2gray(apple); % convert image to grayscale
level = graythresh(grayapple); % get threshold value
BWapple = imbinarize(grayapple,level); % binarize image using the threshold value
BWopen = bwpropfilt(~BWapple,"area",1,"largest"); % filter only the 1 largest blob. The ~ inverts the mask
BWfilled = imfill(BWopen,"holes"); % fill any missing holes
finalimage = imoverlay(apple,~BWfilled,"k"); % burn the mask onto the original image, with black (k) background
imshow(finalimage) % show the final image

MATLAB: Add a field to structure

MaskIm = imread('text.png');
Sat1=regionprops(MaskIm, {'Area', 'Eccentricity', 'EulerNumber', 'MajorAxisLength', 'MinorAxisLength', 'Orientation', 'Perimeter'});
AR = [Sat1.MajorAxisLength]./[Sat1.MinorAxisLength] ;
[Sat1.AspectRatio] = deal(AR) ;

Best Answer

Related Solutions

MATLAB: Create mask that separates objects from white background

MATLAB: Add a field to structure

Related Question