How Can I detect a blurred line out of a low resolution image in Matlab

This is an IR frame that comes from an IR video. I took this video from an orchard with an IR camera mounted on a drone. I need to find the white pipe in each of the frame. But I have to adjust the "houghpeaks" and "im2bw" level parameters every time to have the pipe detected in different frame. Also I need the white pipe to be detected as a continuous one from the left of the image all the way to the right side. But even if the pipe was detected, it has several blind spots due to the branches. It really took me forever to get this far, and don't have any ideas of how to finish this. The code I use is:

 close all
 clear
 clc

 % Read in image and the image info
 I = imread('0003frame103.jpg');
 info = imfinfo('0003frame103.jpg', 'jpg');

 % Convert to binary image
 Level = 0.57;
 Bimage = im2bw(I,Level);

 [H, T, R] = hough(Bimage);

 figure, imshow(imadjust(rescale(H)),[],...
       'XData',T,...
       'YData',R,...
       'InitialMagnification','fit');
 xlabel('\theta (degrees)')
 ylabel('\rho')
 axis on
 axis normal 
 hold on
 colormap(gca,hot)

 % Find the peaks in the Hough transform matrix, H, using the houghpeaks function.
 P = houghpeaks(H,5,'threshold',ceil(0.3*max(H(:))));

 % Superimpose a plot on the image of the transform that identifies the peaks.
 x = T(P(:,2));
 y = R(P(:,1));
 plot(x,y,'s','color','black');

 lines = houghlines(Bimage,T,R,P,'FillGap',5,'MinLength',7);

 figure
 imshowpair(I,Bimage,'montage'), 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','red');

And the image is attached.

Thank you so much for your time and help

Answer 1

[Matt Sprague]

I was able to get good results on this image from using the horizontal components of the Sobel edge filter before inputting to your current workflow using the Hough transform. However I am doubtful that this particular code would be robust for the other images collected. Further adjustments to the threshold used in the "edge" command and filtering of the hough lines may be needed. More info on the edge can found be found through this link.

https://www.mathworks.com/help/images/ref/edge.html

close all
clear
clc
% Read in image and the image info
I = imread('0004frame153.jpg');
I = I(:,:,1);

%Sobel Edge filter
BW = edge(I,'Sobel',0.3,'horizontal');

[H,theta,rho] = hough(BW);

P = houghpeaks(H,3,'threshold',ceil(0.3*max(H(:))));
x = theta(P(:,2));
y = rho(P(:,1));
lines = houghlines(BW,theta,rho,P,'FillGap',5,'MinLength',7);
imshow(BW)

figure, imshow(I), hold on

%Extract endpoints from individual hough lines
points = [];
for k = 1:length(lines)
    points = [points; lines(k).point1; lines(k).point2];
end

% Fit overall line from individual lines from hough
x1 = 1:size(I,2);
p = polyfit(points(:,1),points(:,2),1);
y1 = polyval(p,x1);

plot(x1,y1,'r','LineWidth',3)