This repository accompanies the laboratory practice on Planar Distance Estimation for the AI4Automotive course at University of Modena and Reggio Emilia. Please note that this implementation has only a didactic purpose.
Qualitative results. Whole video here.
The code entry point is main.py.
Please notice that to reproduce results in the demo the following things are needed:
- Videos from the DR(eye)VE dataset. The demo is run on sequence 06. The code expects to find unrolled frames in
data/frames. One example frame is already indata. - Homography matrix to warp the road plane from frontal to bird's eye view. Example matrix is included in
data. - YOLOv3 detector pre-trained weights (downloadable here).
The PyTorch wrapper to YOLOv3 detector is borrowed from this clean implementation by eriklindernoren. Thanks Erik! :)
Slides of the practice are available here.