Hough transform doesn't detect some lines

Hi, I tried using Hough transform to detect all the straight lines in an image below

so that only logic gates remain in the image. But Hough transform detect only some lines as shown in figure an the detected lines are colored in green.

Can someone please tell what could the reason. What is the best, robust function in matlab to detect lines in image that can work on any type of image. If Hough transform is the best available one, what can be done to increase its robustness to detect all straight lines that can be used in any type of image.The code used is below one where BW_ConnComp is a binary inverted image. The "lines" calculated is drawn in green.

img = edge(BW_ConnComp,'prewitt');
figure, imshow(img), hold on
[H,T,R] = hough(img);
P  = houghpeaks(H,5,'threshold',ceil(0.3*max(H(:))));
lines = houghlines(BW,T,R,P,'FillGap',5);

[Jack Smith]

Tried increasing threshold value. Still unable to detect all the lines * (see my comment to first answer below) * . Someone please help.

[Jack Smith]

Also please suggest how to find total number of separate line segments detected.

Answer 1

[Image Analyst]

Try changing the threshold, or not calling edge() at all. Not sure why you called edge in the fist place. I mean it already has edges and calling edge just turns a single edge into a double edge - just look at your image and you'll see.

[Jack Smith]

Thank you for the answer. I tried increasing the threshold value from 0.3*max(H(:)) to 0.5*max(H(:)) and removed edge() function. I only got an improvement of detection of just one more line (line B/W NOT & AND gates) , and still three lines are yet to be detected (as in above fig four lines are undetected). If I try to increase threshold value further, then the lines already detected are also getting undetected.

Answer 2

[Image Analyst]

Jack:

Try the attached code. Change your folder and image name before you run it.

[Walter Roberson]

Jack Smith commented

I want to understand why Hough transform is unable to detect all the lines.

[Image Analyst]

Jack, they were probably not selected due to the threshold level. Attach a specific example (code plus image) in a new question if you want more help.

Answer 3

[sonia carole]

I = imread('SAM_0160.jpg');
I=imresize(I,[640 480]);
figure,imshow(I);
rotI =imrotate(I,33,'crop');
bw_I =rgb2gray(rotI);
BW = edge(bw_I,'canny');
figure; imshow(BW);
[H,T,R] = hough(BW);
imshow(H,[],'XData',T,'YData',R,...
          'InitialMagnification','fit');
xlabel('\theta'), ylabel('\rho');
axis on, axis normal, hold on;
P  = houghpeaks(H,5,'threshold',ceil(0.6*max(H(:))));
x = T(P(:,2)); y = R(P(:,1));
plot(x,y,'s','color','white');
% Find lines and plot them
lines = houghlines(BW,T,R,P,'FillGap',30,'MinLength',15);
figure, imshow(rotI), hold on
max_len = 0;
for k = 1:length(lines)
 xy = [lines(k).point1; lines(k).point2];
 plot(xy(:,1),xy(:,2),'LineWidth',2,'Color','green');

   % Plot beginnings and ends of lines
   %plot(xy(1,1),xy(1,2),'x','LineWidth',2,'Color','yellow');
   %plot(xy(2,1),xy(2,2),'x','LineWidth',2,'Color','red');

   % Determine the endpoints of the longest line segment
   len = norm(lines(k).point1 - lines(k).point2);
   if ( len > max_len)
      max_len = len;
      xy_long = xy;
   end
end

% highlight the longest line segment %plot(xy_long(:,1),xy_long(:,2),'LineWidth',2,'Color','blue');

[sonia carole]

I am facing the same problem,I am not able to detect all the lines in my image, when I am changing both angle and threshold. thank you for your help. my intentions is to detect overhead power lines and then track them.

[Spencer Glubay]

I think your min length is too high or you are don't have enough number of peaks. Both of these will reduce how many lines are displayed in the end.

[Image Analyst]

Try taking the radon transform and see if you can see peaks at the expected angles.

Answer 4

[Satadru Mukherjee]

Simple Code , no need Hough--

clc
clear all
close all
warning off
x=imbinarize(rgb2gray(imread('Capture.JPG')));
imshow(x);
impixelinfo;
[r c]=size(x);
l=zeros(r,c);
se=strel('line',60,0);
imshow(x);
l=l+imopen(x,se);
imshow(l);
se=strel('line',60,90);
imshow(x);
l=l+imopen(x,se);
imshow(l);