Abstract:
Autonomous driving and ADAS tasks are growing. However, porting multiple functions into a power-constrained computing device is difficult. This work aims to simplify the pixel-wise driving scene understanding learning process. This paper uses semantic segmentation as a point detection task to detect free space and lanes. Instead of pixel-wise learning, we trained a single deep convolution neural network for point of interest detection in a grid-based level and post-processed end branches with target class characteristics using computer vision (CV). We propose CV-based post-processing to decode neural network output points to achieve pixel-wise semantic segmentation and parametric lanes. The final results showed that the network could learn the spatial relationship for point of interest, including representative points on the contour of the free space segmented region and representative points along the road lane center. On two publicly available datasets, our method achieved 98.2% mIoU on the KITTI dataset for free space evaluation and 97.8% accuracy on the TuSimple dataset (with the field of view below the y=320 axis) for lane evaluation.
Note: Please discuss with our team before submitting this abstract to the college. This Abstract or Synopsis varies based on student project requirements.
Did you like this final year project?
To download this project Code with thesis report and project training... Click Here
