How to localize lane markings of road in different light conditions

Hi everyone; I am having two pictures of the same road taken during different light conditions as shown at: http://tinypic.com/view.php?pic=rjmupy&s=6 The first picture is taken in evening and the second one during daylight. I have to write one general code which determines the white lane marks of the road for both light scenerios. I have tried ways like sobel edge detection, connected components and have also tried to find the white lane marks by applying conditions on white pixel values.But the algorithm which works on the first picture doesnot work on the second one. Please guide me.

Answer 1

[Image Analyst]

We can't help you perfect your road following algorithm. I suggest you look at papers of those who have done it successfully and published their results here: http://iris.usc.edu/Vision-Notes/bibliography/contentsactive.html#Active%20Vision,%20Camera%20Calibration,%20Mobile%20Robots,%20Navigation,%20Road%20Following, especially this section:

15.3.3.2 Lane Detection, Lane Following, White Line Detection

The Mathworks also has an example in their Computer Vision System Toolbox: http://www.mathworks.com/help/vision/examples/lane-departure-warning-system-1.html?prodcode=VP&language=en