An application with a user-friendly interface, easy-to-use using OpenCV, YoloV3 to detect and count the number of vehicles passed through a line in the video, live cam, or stream. This app runs in multi-threading for different videos and streams.
Note: You should have a GPU with NVIDIA driver, CUDA toolkits and CuDNN to run faster with YoloV3
git clone https://github.com/binh234/vevhicle_couting.git
To install required dependencies run:
pip install -r requirements.txtpython main.pyGUI: github.com/flytocc/pyqt5-cv2-multithreaded
YoloV3: pjreddie.com/darknet/yolo/
SORT: github.com/abewley/sort
@inproceedings{Bewley2016_sort,
author={Bewley, Alex and Ge, Zongyuan and Ott, Lionel and Ramos, Fabio and Upcroft, Ben},
booktitle={2016 IEEE International Conference on Image Processing (ICIP)},
title={Simple online and realtime tracking},
year={2016},
pages={3464-3468},
keywords={Benchmark testing;Complexity theory;Detectors;Kalman filters;Target tracking;Visualization;Computer Vision;Data Association;Detection;Multiple Object Tracking},
doi={10.1109/ICIP.2016.7533003}
}