Modern lane detection algorithms outperform traditional lane detectors by extracting and predicting lane features using a range of deep learning approaches. Deep learning algorithms, however, are computationally taxing and frequently fall short of the real-time demands of autonomous cars. This study suggests a lane detection technique that takes advantage of deep learning's potential while addressing real-time requirements. It uses a lightweight convolutional neural network model as a feature extractor. The created model is trained on a dataset of 16 x 64 pixel tiny image patches, and fast inference is accomplished by using a non-overlapping sliding window technique. Following that, a polynomial is fitted to the predictions in order to model the lane boundaries. On the KITTI and Caltech datasets, the proposed technique was evaluated, and it showed acceptable performance. We also incorporated the detector into our autonomous vehicle's localization and planning system. It operates at 28 frames per second in a CPU with an image resolution of 768 x 1024, which satisfies the real-time demands necessary for self-driving automobiles. final year project is based on IEEE Paper.this will be one of the best Final year engineering project for computer science. Components that we will provide are.
1.complete documentation support
2.complete working hardware/software implemented in students environment
3.classes will held accordingly.