Lane-Detection

This project is implemented to detect lane and predict the turn of the given video. The pipeline is as follows.

• The video is taken frame by frame.
• The frame is undistorted using camera parameters.
• Noise in the frame is removed using Gaussian filter.
• Bottom half of the frame is taken as Region of Interest.
• Homography is applied to the frame to get the bird's view of the road.
• Histogram is taken to detect the lane lines in the road.
• Lane lines are filtered using color segmentation.
• A polynomial is fitted to both right and left line of the road.
• The polynomial is superimposed back to the original frame using inverse Homography.

Results

Run Instruction

1. Run the Lane_Detection.py.
2. Select the video by entering either 1 or 2.