Engineers Garage

Matlab Red Color Detection

By

Description

This article explains how to detect and track red objects in a live video. There are different approaches to detect the red color in every single frame. One of these popular approaches includes the conversion of whole RGB frame into corresponding HSV (Hue-Saturation-Value) plane followed by extraction of pixel values only for red. One can choose a range that reflects all existing shades of red so as to detect and differentiate all colors in the frame.
 
However, this approach is practically difficult to apply especially in Live video due to the ambient light. There is one more solution where one can detect a particular color say red, blue or green.   Although this is not versatile for all colors, but it easily eliminates the issue of ambient light. Here, we are going to use the same approach in order to detect red color. 
 
For newbie people
 
“ Do coding into ‘editor’ of Matlab “
 
Flow diagram 
 

Flow Diagram of Matlab Code used for Color Detection

Fig. 1: Flow Diagram of Matlab Code used for Color Detection

Steps Required

Step 1: First acquire an RGB Frame from the Video.
Matlab code :

Image showing blocks of different shapes in red color

Fig. 2: Image showing blocks of different shapes in red color

A=imread(a.jpg);
 
 
 
Step 2: Extract the Red Layer Matrix from the RGB frame.
MATLAB Code:

Image showing extracting red layer matrix

Fig. 3: Image showing extracting red layer matrix

redA = A(:,:,1);
Step 3: Get the grey image of the RGB frame.
MATLAB Code:

Image showing transformation of captured image after adding grey frame

Fig. 4: Image showing transformation of captured image after adding grey frame

grayFrame = rgb2gray(A);
Step 4: Subtract the grayFrame from the redFrame.
MATLAB Code: 

Image showing transformation of captured image after subtracting grey frame from red matrix

Fig. 5: Image showing transformation of captured image after substracting grey frame from red matrix

diffFrame=imsubtract(redA, grayFrame);

Step 5: Filter out unwanted noises using Median Filter
MATLAB Code: 

Image showing transformation of captured image after filtering out unwanted noises

Fig. 6: Image showing transformation of captured image after filtering out unwanted noises

diffFrame = medfilt2(diffFrame, [3 3]);
Step 6: Now convert the diffFrame into corresponding Binary Image using proper threshold value. Change its value for different light conditions. In my code I have used its value as 0.15.
MATLAB Code:

Image showing final binary image

Fig. 7: Image showing final binary image

binFrame = im2bw(diffFrame, 0.15);

Step 7: Now your Red color has been detected. You can put any blob statistics analysis on this image or calculate the centroid, area or bounding box of those blobs.
I have introduced the same algorithm in the code of this project that you can see below. 
 
Software & hardare Required

1 Matlab (here I am using R2011 b)

2 Webcam ( inbulid or external)

Screenshot of Matlab Code used for Red Color Detection

Fig. 8: Screenshot of Matlab Code used for Red Color Detection

 

Screenshot of Matlab Command Window implementing color detection program

Fig. 9: Screenshot of Matlab Command Window implementing color detection program

 

Project Source Code

###
%%take video from webcam
vid=videoinput('Winvideo',1,'YUY2_320x240'); 
%set frames trigger
set(vid, 'FramesPerTrigger', Inf); 
% video output is rgb color
set(vid,'ReturnedColorspace','rgb';
% set video frame interval 1
vid.FrameGrabinterval=1
% start video
start(vid);
%start loop till frame 10000
while(vid.FramesAcquired<=1000)
% get snapshot of video
data=getsnapshot(vid);
% red color extraction
red=data(:,:,1);
% convert picture into gray
gray=rgb2gray(data);
% subtract red to gray frame
diff_im=imsubtract(red,gray);
% use median filter
diff_im=medfilt2(diff_im,[3,3]);
% convert image into binary
diff_im=im2bw(diff_im,0.10);
% use blob statistics analysis on this image 
diff_im=bwareaopen(diff_im,1000);
bw=bwlabel(diff_im,8);
% use boundbox to enbox red color object
stats=regionprops(bw,'BoundingBox','Centroid');
image(data);
 
hold on
% identify no of red object in image
for(object=1:length(stats))
bb=stats(object).BoundingBox;
bc=stats(object).Centroid;
 
rectangle('Position',bb,'EdgeColor','r','LineWidth',2)
plot(bc(1),bc(2),'-m+')
a=text(bc(1)+15,bc(2),strcat('X: ',num2str(round(bc(1))),' Y: ',num2str(round(bc(2)))));
set(a,'FontName','Arial','FontWeight','bold','FontSize',12,'Color','yellow');
if(length(stats)==1)
% show label one with one object
a=text(50,60,strcat('One'));
set(a,'FontName','Arial','FontWeight','bold','FontSize',12,'Color','red');
end
if(length(stats)==2)
% show label two with two red object 
a=text(50,60,strcat('Two'));
set(a,'FontName','Arial','FontWeight','bold','FontSize',12,'Color','red');
end
if(length(stats)==3)
% show label three with three red object
a=text(50,60,strcat('three'));
set(a,'FontName','Arial','FontWeight','bold','FontSize',12,'Color','red');
end
if(length(stats)==4)
% show label four with four red object
a=text(50,60,strcat('four'));
set(a,'FontName','Arial','FontWeight','bold','FontSize',12,'Color','red');
end
end
hold off
% end loop 
end
% stop video
stop(vid);
% flush data
flushdata(vid);
% clear all


clear all; 



###
 

Project Video

Filed Under: Electronic Projects

 

Questions related to this article?
👉Ask and discuss on EDAboard.com and Electro-Tech-Online.com forums.

Tell Us What You Think!!

Search Engineers Garage