How to use threshold in this problem?

I convert this picture into bw using c=im2bw(b,graythresh(b)); but the problem is the straight line is disconnected as picture below. It supposed to be straight single line. What kind of segmentation technique that I need to use?

Answer 1

[Alessandro Masullo]

Supposed your image varies from o to 1, you could use a threshold in this way:

   THR = 0.5 
   above = image>=THR;
   image(above) = 1;
   image(~above) = 0;

[muhammad ismail]

So how to apply in this coding?

a=imread('S2a.jpg');

b= imcrop(a,[83 48 550 353]);

c=im2bw(b,graythresh(b));

[Alessandro Masullo]

a=imread('S2a.jpg');
b= imcrop(a,[83 48 550 353]);
c = im2double(b);
THR = 0.5 
above = c>=THR;
c(above) = 1;
c(~above) = 0;

[muhammad ismail]

But it display like picture below. a=imread('S2a.jpg');

b= imcrop(a,[83 48 550 353]);

c = im2double(b);

THR = 0.5;

above = c>=THR;

c(above) = 1;

c(~above) = 0;

imshow(c);

[Alessandro Masullo]

That's because your line gray level is the same of the background, so you cannot use a simple conversion from gray to b/w, but you have to use an edge detection algorithm, as Image Analyst suggested:

[muhammad ismail]

Ok now I understood. Thank you very much for your great explanation. By the way I have last question for you. How to display gray level like image above. Sorry I'm newbie in matlab.

[Alessandro Masullo]

I just used imagesc(c). You can use different colormaps instead of the grey levels, use:

   colormap jet
   help colormap

Answer 2

[Image Analyst]

Like I said in your duplicate post, you can use imclearborder() and bwareaopen(), though there are better ways like background correction.