This project aims at estimating traffic flow from CCTV footage. We have trained YOLOv4 for the following five vehicle classes.
- car
- bus
- truck
- van
- bike
We obtained the images through the traffic cameras API, New Zealand Transport Agency (NZTA). The image dataset with the relevant annotation files used to train YOLOv4 is available in the "traffic-detection-dataset" directory. We trained YOLOv4 with nearly 16, 100 vehicle object instances as illustrated in the following table.
| Vehicle Class | Total Instances |
|---|---|
| Car | 3,627 |
| Bus | 141 |
| Van | 779 |
| Truck | 1,273 |
| Bike | 280 |
Our experimets were carried out for a footage we received from NZTA for a busy mult-lane road in Christchurch CBD.
