MATLAB: Extract Shapes (Polygons) from boundaries

MATLABshape polygon boundary

I have implemented code in Matlab to analyze an image of a PCB and then highlight possible ICs via edge detection.

Is there any way for me to use my set of boundaries (shown in red in the attached image) to extract sub-images of board components from the master images? If possible, I would only like to derive and then extract square and rectangular shaped areas so I skip board silkscreen and holes.

Best Answer

I doubt you'll find ready-made code to do it. However, you can use houghlines() to fit line segments to your edge points. Once you do, you can test which line segments are perpendicular to each other, etc...

Related Solutions

MATLAB: Detect and Fill Edges of a Binary Image

Assuming that the inner/outer ring can be approximated by the edge of convex hull of inner/outer ring points, the area between inner and outer ring can be extracted by the following:

% Read and prepare the binary image
I = imread('Slice.jpg');
Igray = rgb2gray(I);
BWin = imbinarize(Igray);
BWin = imclearborder(BWin);
% Convex hull of outer ring points
CH1 = bwconvhull(BWin);
% Remove outer ring points
se = strel('disk',10);
CH1b = imerode(CH1,se);
BW1 = BWin & CH1b;
% Convex hull of inner ring points
CH2 = bwconvhull(BW1);
% Extract the area between inner and outer convex hull
BWout = CH1 - CH2;
% Show the result
figure
imshowpair(BWin,BWout)

MATLAB: Additional properties from regionprops

You could also return the "ConvexHull" values which will return a p-by-2 matrix for each polygon where each row of the matrix contains the x- and y-coordinates of one vertex of the polygon. Then you could use pdist() to compute the length of each segment.

Best Answer

Related Solutions

MATLAB: Detect and Fill Edges of a Binary Image

MATLAB: Additional properties from regionprops

Related Question