How to remove time lag and multiple centroid in this code?

The cursor does not follows red color properly and it detects white and various other color also. currently i am doing virtual mouse using color detection. this is code please help.
clear all
close all
clc
import java.awt.Robot
m=Robot;
v=videoinput('winvideo',1,'YUY2_640x480');
set(v,'ReturnedColorSpace','YCbCr');
v.FramesPerTrigger=inf;
v.FrameGrabInterval=1;
pause(1)
while (v.FramesAcquired<=35)
i=getsnapshot(v);
% i=imresize(i,2);
% imtool(im_resized)
first_comp=i(:,:,1);
% imtool(first_comp)
rgb_2_gray=rgb2gray(i);
% imtool(rgb_2_gray)
d=first_comp-rgb_2_gray;
% imtool(d)
b=im2bw(d,.1);
bw=bwareaopen(b,500);
imshow(bw);
hold on;
% end;
% % bw=bwlabel(bw,8);
% % imtool(bw)
p=regionprops(bw,'centroid');
for z=1:length(p)
a=p(z).Centroid;
plot(a(1),a(2),'r*');
m.mouseMove(a(1),a(2));
end;
end;

1 Kommentar

For your reference, your code is now much more readable because I applied this formatting button to it,

Melden Sie sich an, um zu kommentieren.

 Akzeptierte Antwort

Image Analyst
Image Analyst am 26 Sep. 2014

0 Stimmen

It's just not a very robust color segmentation algorithm. You might be better off going to hsv color space. See my File Exchange. However one of the methods I use there does find red objects in RGB color space. http://www.mathworks.com/matlabcentral/fileexchange/26420-simplecolordetection

2 Kommentare

This link was very helpful to understand quickly now will be able to detect green and blue color and perform clicking accordingly.
I just looked at it. It's not any more robust than yours was - in fact it's pretty much the same algorithm. Well, whatever, if it works in your specific situation, then fine.

Melden Sie sich an, um zu kommentieren.

Weitere Antworten (1)

Rohit Dixit
Rohit Dixit am 27 Sep. 2014

0 Stimmen

Thank you sir good to know. I will improve the robustness in code.

Community Treasure Hunt

Find the treasures in MATLAB Central and discover how the community can help you!

Start Hunting!

Translated by