Red Ball Tracking Problem

Hi. I have a question. A maked project wich should track a red ball. But appear the problem which bedevil me. There i put my code , if someone can help me, please.

if true
 imaqhwinfo
dev_info = imaqhwinfo('winvideo',1)
%info=imaqhwinfo('winvideo')

   celldisp(dev_info.SupportedFormats)
  vid=videoinput('winvideo',1,'YUY2_640x480');
%%Set the properties of the video object
set(vid, 'FramesPerTrigger', Inf);
set(vid, 'ReturnedColorspace', 'rgb')
vid.FrameGrabInterval = 5;
%Achizitie video
start(vid)
%Initializare arduino
ser=serial('COM3','Baudrate',9600);
fopen(ser)

while(vid.FramesAcquired<150) data = getsnapshot(vid); %din gray extragem componenta rosie r = data(:,:,1);g = data(:,:,2); b = data(:,:,3); bw1 = ((r >=248) & (g <=2) & (b <=2)); %bw1 = (r - g/2 - b/2); % % stergem pixelii mai mici deci 30 bw1 = bwareaopen(bw1,30); bw = imfill(bw1,'holes'); [B,L] = bwboundaries(bw,'noholes'); %figure(1); subplot(224); imshow(bw1); %nr de componente conectate din imagine %bw = bwlabel(bw1, 8); % bw1 = imfill(bw1,'holes'); % [B,L] = bwboundaries(bw,'noholes'); %Blob analiza stats = regionprops(L,'Area','Centroid'); circle_threshold = 0.80; figure(1); imshow(data); hold on for g = 1:length(B)

          % obtain (X,Y) boundary coordinates corresponding to label 'g'
          boundary = B{g};

          % compute a simple estimate of the object's perimeter
          delta_sq = diff(boundary).^2;
          perimeter = sum(sqrt(sum(delta_sq,2)));

          % obtain the area calculation corresponding to label 'g'
          area = stats(g).Area;

          % compute the roundness metric
          metric = 4*pi*area/perimeter^2;
          if metric > circle_threshold
               bc = stats(g).Centroid;
               plot(bc(1),bc(2),'ko');
               a=text(bc(1)+16,bc(2),strcat('X:   ',num2str(round(bc(1))),'    Y:  ', num2str(round(bc(2)))));
               set(a, 'FontName', 'Arial', 'FontWeight','bold','FontSize',12,'Color','yellow');
       %Poz=uint8(bc(1));

        X = round(bc(1));
%         valoarea_prag = 60;
%         Xc = 320;%320;
      [pancmd] = pantilt(X);
       fwrite(ser,pancmd,'char');
%         if (Xc-X) > valoarea_prag 
%         fprintf(ser,'r');
%           %%Find out if the X component of the face is to the right of the middle of the screen.
%          elseif (X-Xc) > valoarea_prag
%           fprintf(ser,'l');
%           end  
       else
                  [r,c] = find(L(:,:)==g);
                % [r,c] = find(bwlabel(bw)==g);
                  bw(r,c)=0;
            end
   end

   %  pause
    hold off
end
%b.servoWrite(9,0);
%pos = b.servoread(1);

%oprire achizite video closepreview(vid) stop(vid); %delete(vid)

fclose(ser) %sterge toate variabilele clear all sprintf('%s','gata') end

Answer 1

[Image Analyst]

See my attached code where I track a green object. Feel free to change the values to track a red object instead.

[Stas Radion]

Thank You for code! But still want to know what is wrong in code. To know on the future

[Image Analyst]

I don't know. I didn't run it, and you didn't say what was wrong. Is there an error message? Does it not find the red region(s)? Try this: http://blogs.mathworks.com/videos/2012/07/03/debugging-in-matlab/

[Stas Radion]

i make

if true
 bw1 = ((r >=248) & (g <=100) & (b <=100));
end  
and he recognize one type of red color. Now he show me many coordonation in image , how can i do to make just one X

 .

[Image Analyst]

That should show a whole range of colors. If it's not doing the job, try using delta E like I did in my demo: http://www.mathworks.com/matlabcentral/fileexchange/31118-color-segmentation-by-delta-e-color-difference

[Stas Radion]

Thank You for suggestion. I will try this.

Answer 2

[Syed Asad Ali]

Hi..sir please understand me about this intensity formula and threshold for finding green intensity....

r=center(:,:,1); g=center(:,:,2); b=center(:,:,3); green=g-(r/2)-(b/2); bw=(green>30);

in this formula why use /2...and why g is place before r/2 and b/2..and also why use green>30...why use 30 in this...sir pls ans ASAP

[Image Analyst]

I answered in your duplicate question here.