Host Lane Detection Method Using Semantic Segmentation Combined with Hierarchy Clustering Algorithm

Lane detection is the core module of autonomous cars and plays an important role in lane keeping and trajectory planning. Traditional works on lane detection depend on hand-crafted features and are weak to road scene variations. Recent end-to-end approaches leverage deep learning models, but are computationally demanding and need heavy work of labeling lanes. In this paper, we combine the advantage of traditional method and deep learning based method. Our lane detection method has three key novelties: (1) we use semantic segmentation based deep learning to extract region of interest (ROI), which can free the work of manually selecting features and fast locate the lane area. (2) We propose a lane feature points extracting algorithm based hierarchical clustering to effectively remove the disturbance of the noise. (3) we make use of the similarity of the inter-frame to correct the lane fitting results. Experimental results show that our lane detection method can adapt to various road scene and significantly decrease the false positive rate.