Self-Driving - Lane Detection: program to detect lanes in a video

Lane Detection on National Highway:

Version 1.0:

• Added smoothness of the average line between frames, to make the average line much more stable

Version 0.9:

• Added a average of all the lines, as a single line in the video

Version 0.85:

• Increase the quality of lines predicted by dilating the frame after canny function, which increases the width of the foreground objects making it easier to capture lines

Version 0.8:

• Improve lines detection, using improved parameters values for HoughLInesP

Version 0.5:

Every frame from the video:

• is turned into grey-scale version using cv2.cvtColor function
• Finds edges in the frame using the [Canny86] algorithm by computing gradient, to identify change in pixels
• Using region_wants function discards the non-important areas of the frame for lane detection
• cv2.HoughLinesP to detect different lines passing through points detected by canny method
• Draw these lines on blank image of frame's size
• Use cv2.addWeighted function, to combine the original frame and empty image with drawn lines