Asked by john
on 10 May 2013

My research recently is focus on using color stripe to detect 3d model. I generate a color stripe using De Bruijn sequence. But the edge detection is not as good as I expected.I am afraid this image could not be edge detected well, it is the first important step to process! How to get a good image?My aims is to detect the edge between color stripes?

%%initialize

clear;close all;clc;

I = imread('small2.jpg');

I2 = rgb2gray(I);%I2 gray image

I3=I2;

figure(1);

image(I);

axis image;

N=size(I);

%Notice our stripe is vertical, so we check every column to detect edges.

R=I(:,:,1);

G=I(:,:,2);

B=I(:,:,3);

%%edge detection

%threshold

TS=10;

for i=1:N(1) %row

for j=2:N(2)

Defference=abs(R(i,j)-R(i,j-1))+abs(G(i,j)-G(i,j-1))+abs(B(i,j)-B(i,j-1));

if(Defference>TS)

I2(i,j-1)=255;

end

if((i<510)&&(i>498))

I2(i,j-1)=255;

end

end

end

%%get a part of the image and show its RGB content in a graph.

set(1,'color',[1 1 1]);

title(['input 2 points，get a rectangular, show its RGB curve' num2str(1)]);

% Use ginput to select corner points of a rectangular

% region by pointing and clicking the mouse twice

p = ginput(2);

% Get the x and y corner coordinates as integers

sp(1) = min(floor(p(1)), floor(p(2))); %xmin

sp(2) = min(floor(p(3)), floor(p(4))); %ymin

sp(3) = max(ceil(p(1)), ceil(p(2))); %xmax

sp(4) = max(ceil(p(3)), ceil(p(4))); %ymax

% Index into the original image to create the new image

MM = I(sp(2):sp(4), sp(1): sp(3),:);

% Display the subsetted image with appropriate axis ratio

figure(2);

subplot(3,3,1);image(MM); axis image;title('the image extract');

hold on;

WW=40;

COnes=ones(WW);

CZeros=zeros(WW);

Mycolor=[0 0 0;0 0 1;0 1 0; 0 1 1; 1 0 0;1 0 1;1 1 0;1 1 1];

for i=1:8

subplot(3,3,i+1);

image(cat(3,Mycolor(i,1)*COnes,Mycolor(i,2)*COnes,Mycolor(i,3)*COnes));

title(num2str(Mycolor(i,1))+num2str(Mycolor(i,2))+num2str(Mycolor(i,3)));

end

N=size(MM);

RowMax=min(8,N(1));

ColumnMax=min(900,N(2));

x=1:ColumnMax;

figure(3);

for i=1:RowMax

subplot(RowMax,1,i);

y1=MM(i,1:ColumnMax,1);

y2=MM(i,1:ColumnMax,2);

y3=MM(i,1:ColumnMax,3);

plot(x,y1,'r');

hold on;

plot(x,y2,'g');

hold on;

plot(x,y3,'b');

end

%%show the edge detection result in gray mode,edge is showed in red line.

I4=cat(3,I2,I3,I3);

figure(4);

imshow(I4);

You can see in the background there are black lines seems like net. 'Rapid Shape Acquisition Using Color Structured Light and Multi-pass Dynamic Programming'.is the paper I based on.

This picture is I taken 2 days ago.My edge detection is totally failed. The key to success is to find each color transition through edge detection. But the line in background and the color cross talk is very serious!! Thank you very much.

It uses edge detection to find each color transition line.But it won't work well I guess.

Opportunities for recent engineering grads.

Apply Today
## 3 Comments

## Jaco Verster (view profile)

Direct link to this comment:https://www.mathworks.com/matlabcentral/answers/75287-does-anyone-do-research-on-structured-light-how-to-get-a-good-edge-detection#comment_400842

## Tkingdom (view profile)

Direct link to this comment:https://www.mathworks.com/matlabcentral/answers/75287-does-anyone-do-research-on-structured-light-how-to-get-a-good-edge-detection#comment_411687

## Camilo (view profile)

Direct link to this comment:https://www.mathworks.com/matlabcentral/answers/75287-does-anyone-do-research-on-structured-light-how-to-get-a-good-edge-detection#comment_599077

Sign in to comment.