Key points are not available for this paper at this time.
An improved vision-based framework is proposed for lane detection and departure warning system. It consists of an edge filter developed based on sliding discrete cosine transform for lane features extraction, RANdom SAmple Consensus (RANSAC) method for model fitting of the lanes, and a lateral offset ratio computation on lanes for warning system. A finding of edge detection kernel extracted from 2-D SDCT is reported with its evaluation on road images. FPGA implemented edge detector's performance is proved to be better compared to existing edge detectors in terms of Pratt's figure of merit, average gradient, and entropy for road images. Moreover, the edge detection capability under noisy condition of 10 dB SNR is proved to be the best. The lane detection and warning system are implemented in real-time processor of compact reconfigurable input output system using vision development module. Lane departure warning is issued based on the computed LOR using X-coordinates of detected bottom end points of lane lines in an image plane. The lane detection performance is evaluated on the Caltech lane dataset, KIST lane dataset, and LaneVisionIITR dataset with diversity of 4329 images altogether. The proposed framework shows the performance improvement on urban and highway driving in complex scenarios.
Sukumar et al. (Thu,) studied this question.
Synapse has enriched 5 closely related papers on similar clinical questions. Consider them for comparative context: