This repository consists of the code files used in the UG Project done under Dr. Sobhita Mehar, Assistant Professor, Department of Electrical Engineering.
A three-way approach was followed to complete this lane detection project. To understand the core fundamentals of detection in an image, firstly, a foundation approach was selected, which uses canny detection and Hough transform algorithm with some preprocessing to detect lanes. After that, a more optimized approach was discussed, using dual light sensors to enhance the quality of lane detection in critical conditions, such as tunnel and twilight driving. Once these fundamentals were built, a more advanced approach with deep learning was selected, which utilizes the fully connected CNN pre-trained model to do prediction on the lane.