segmentation of rice grains and sub plotting individual grains

in the code attached, i want all the segmented rice images to appear in the form of subplot but this code keeps me giving the output as the last rice image in subplot.Thanks in advance.(this code was developed with the help of some codes present on the internet)

a=imread('r-1121.jpeg');

figure,imshow(a);

title('original image');

I1 = rgb2gray(a);

I1=imsharpen(I1);

I=medfilt2(I1);

b=imbinarize(I);

figure,

imshow(b);

c=imfill(b,'holes');

label=bwlabel(c);

for j=1:max(max(label))

[row,col]=find(label==j);

len=max(row)-min(row)+2;

breadth=max(col)-min(col)+2;

target=uint8(zeros([len breadth] ));

sy=min(col)-1;

sx=min(row)-1;

for i=1:size(row,1)

x=row(i,1)-sx;

y=col(i,1)-sy;

target(x,y)=a(row(i,1),col(i,1));

end

for k=1:25

mytitle=strcat('Rice number: ',num2str(j));

ax=subplot(1,25,k);

imshow(target);

end

end

Answer 1

[DGM]

You don't need the inner plot loop, and some other stuff can be simplified. I just removed the figure and imshow previews for brevity.

a=imread('rice.jpg');
I1 = rgb2gray(a);
I1=imsharpen(I1);
I=medfilt2(I1);
b=imbinarize(I);
c=imfill(b,'holes');

label=bwlabel(c);
numgrains = max(label(:));
for j=1:numgrains
    [row,col]=find(label==j);
    len=max(row)-min(row)+2;
    breadth=max(col)-min(col)+2;
    
    xrange = min(col)-1+(0:breadth);
    yrange = min(row)-1+(0:len);
    target = a(yrange,xrange) .* uint8(c(yrange,xrange)); % NOTE
    
    mytitle=strcat('Rice number: ',num2str(j)); % not used yet
    ax=subplot(1,numgrains,j);
    imshow(target);
end

Alternatively, you can get the location using regionprops():

a=imread('rice.jpg');
I1 = rgb2gray(a);
I1=imsharpen(I1);
I=medfilt2(I1);
b=imbinarize(I);
c=imfill(b,'holes');

S=regionprops(c,'boundingbox');
numgrains = numel(S);
for j=1:numgrains
    tp = round(S(j).BoundingBox);
    xrange = tp(1)+(-1:tp(3)+1);
    yrange = tp(2)+(-1:tp(4)+1);
    target = a(yrange,xrange) .* uint8(c(yrange,xrange)); % NOTE
    
    mytitle=strcat('Rice number: ',num2str(j)); % not used yet
    ax=subplot(1,numgrains,j);
    imshow(target);
end

You might need to tweak the subscript vectors by a pixel to get your padding exactly as you want it.

In both examples, I made a note on the same line. In your original code, the pointwise method only extracts the portion of the image selected by the mask c. If on the other hand, you wanted to select the region of the image defined by the bounding box you calculate around the image, then change that line to

target = a(yrange,xrange);

The first method truncates the edges of the grains according to the mask and makes them jagged. The second method shows the grains with their local background and natural edges. Which to use depends what your intent is for using the images of the objects.

[DGM]

You can try something like this

a=imread('rice.jpg');
I1 = rgb2gray(a);
I1=imsharpen(I1);
I=medfilt2(I1);
b=imbinarize(I);
c=imfill(b,'holes');

S=regionprops(c,'boundingbox');
bb = vertcat(S.BoundingBox); 
maxbb = max(bb(:,[4 3])); % minimal box geometry
maxbb = maxbb+5; % add some extra room

numgrains = numel(S);
for j=1:numgrains
    tp = round(S(j).BoundingBox);
    xrange = tp(1)+(-1:tp(3)+1);
    yrange = tp(2)+(-1:tp(4)+1);
    
    % this was from the prior answer
    %target = a(yrange,xrange) .* uint8(c(yrange,xrange));
    
    % alternatively, start out with no masking
    target = a(yrange,xrange);
    % maybe you could improve the range (normalize)
    % this helps get the photo bg closer to black
    target = mat2gray(target);
    % you could use a dilated mask 
    tm = double(imdilate(c(yrange,xrange),ones(3)));
    % or maybe soften the mask as well
    tm = imfilter(tm,fspecial('gaussian'));
    % apply the mask
    target = im2uint8(im2double(target) .* tm);
    
    % this pads with black to match sizes
    bs = (maxbb-size(target))/2;
    target = padarray(target,floor(bs),0,'pre');
    target = padarray(target,ceil(bs),0,'post');
    
    mytitle=strcat('Rice number: ',num2str(j)); % not used yet
    ax=subplot(1,numgrains,j); 
    imshow(target);
end

I used a 5x5 subplot tiling so it's easier to fit in my window. Use whatever tiling you need.

This could probably be faster if the dilation and blurring were done on a copy of c prior to the loop, but doing it this way makes it easier to demonstrate the sequence of options in one place. I doubt the difference is really worth the extra hassle. Doing the normalization in the loop is preferred, as it only helps to compensate for local lighting variation.

[sana fatima]

Thank you @DGM. will further consullt you if i come across any problem.

Answer 2

[sana fatima]

thank you so much for this code. Had been trying this for so long. just need to ask one thing,

1) can we smooth these jagged boundries(the edges)

2) can the frame size be same for all the grains, some have bigger frame and some have smaller, just need them to be in a frames of same size.

thank you so much once again.

[DGM]

This comment isn't an answer. Just copy this into a comment under the other answer so that this conversation is all in one place. We can talk about it there.

[sana fatima]

should i copy my question above where i had intially posted my question(the first one)?

[DGM]

eh. don't worry about it. i already just posted the followup as a comment. just keep it in mind for the future.

Answer 3

[sana fatima]

I have made a code that gives features like major axis length, minor axis length, ecentricity etc but on average of rice grains present in an image. now i have no idea how should i implement those features on individual rice grains so that when two different varieties of rice grains are given as input it classifies those grains based on their respective features. For suppose when IRRI6 and 386 white rice are given as input in mixed form, it should classify that this grain is of IRRI6 along with its features(major axis, minor axis, ecentricity.Here, I am attaching a screenshot of my output window.