How may i draw a middle line and measure the horizontal detainees between this line and the other dashed two lines

Hi Everyone

     Im trying to write a Matlab code to detect the lans of a highway image. The way that it should work is by drawing line in the middle of the lane and then measure the horizontal distance between the middle line and the right & left dashed lines. This is my command

-------------------------------------

 a  = imread('1.jpg');
        I=rgb2gray(a);
        rn = wiener2(I,[20 20], 0.05);
        d = graythresh(rn); 
        e = im2bw(rn,d); 
        BW  = edge(e,'canny',[0.475,0.575],0.2);   
        BW(1:100, :, :)=0;

        [H,theta,rho] = hough(BW,'Theta',-65:1:65);
        P = houghpeaks(H,15,'threshold',ceil(0.3*max(H(:))));
        lines = houghlines(BW,theta,rho,P,'FillGap',50,'MinLength',1);
        imshow(a), hold on
        max_len = 0;

        for k = 1:length(lines)
        xy = [lines(k).point1; lines(k).point2];
        plot(xy(:,1),xy(:,2),'LineWidth',3,'Color','red');

   L1=plot(xy(1,1),xy(1,2),'x','LineWidth',2,'Color','yellow');
   L2=plot(xy(2,1),xy(2,2),'x','LineWidth',2,'Color','green');

       end
----------------------------------

<<

>>

This is how its look, It doesn't look as i want. I just wanna grow the middle line and measure the diatonic between the middle and the the detected lines( right and left)

Answer 1

[Image Analyst]

I doubt it's going to be as simple as some 50 line long algorithm. Go here to find out the algorithms people have used successfully and published for lane following: http://www.visionbib.com/bibliography/contentsactive.html#Active%20Vision,%20Camera%20Calibration,%20Mobile%20Robots,%20Navigation,%20Road%20Following

15.3.3 Road, Path Following Operators
    15.3.3.1 Road Following, Depth, Stereo Based, Off-Road, Safe Path
    15.3.3.2 Lane Detection, Lane Following, White Line Detection

[Ahmed Alotaibi]

Thanks, is it the only way to do so? or there is an easy why to do it since I'm beginner in image processing tool

[Image Analyst]

A beginner like you might be able to develop an algorithm that works on an image or two but not robust in general. For example, what if your image has yellow lines instead of white? What if the car is in the right lane instead of the left? What if there is snow on the road? What if there are other vehicles in the field of view? What if it's raining? What if it's daylight instead of night? What if it's a 5 lane road instead of a two lane road? What if the lane markers are not that shape, for example they're the short, thick "must exit" lane markers? What if the middle lane marker is a double solid line? What if the middle lane is solid instead of dashed? What happens if there is a cross walk, which can take a huge variety of patterns, in the road? What if there are animals, people, bicycles, etc. in the field of view? Can your short algorithm handle all those cases?

[Ahmed Alotaibi]

That will be perfect if i do one of these.Actually,these questions comes to my mind always however i don't want to make a product. I'm trying to focus on putting the lane in the middle of these two weight lines and measure the distance which is relatively easier than considering other disturbances.could you please narrow my solution!

[Image Analyst]

I don't know what diatonic means. And if you want to find distances between lines, simple algebra and the Pythagorean theorem would work. Not sure what you mean by "grow the line" but simple algebra will let you lengthen a line in the overlay, or if it's part of the image, imdilate will do that.