This project was carried out for educational purpose, its main objective is to apply object detection and tracking on pedestrians and cars from CDW-2012 dataset.
It makes use of OpenCV library to detect objects using Haarcascade or background subtraction and then it tracks the trajectory of each object using different tracking algorithms (opencv-contrib-python
was used because it implements more tracking algorithms compared to opencv-python
).
- Install required python packages.
pip install -r requirements.txt
- Download CDW-2012 dataset and extract dataset folder in the root folder of this project.
- Download GOTURN pretrained model and extract
goturn.caffemodel
in the root folder of this project. - Execute
app.py
.
python app.py