README.md

Volleyball Detection & Tracking

University of Central Asia⛰️

🏃‍♂️ How to Run

1. Clone this repository

git clone https://github.com/shukkkur/VolleyVision.git

2. Install the requirements

cd VolleyVision\Stage I - Volleyball
pip install -r requirements.txt

3. ❗If you are facing errors when trying to use RoboFlow model, please refer to this disscusion

MODELS
"yolov7" - fast | "roboflow" - accurate

Let us test on assets/rally_men.mp4. It's a 5 seconds video that weights about 5.2 MB.

4. If you want to get fast results, use volley_track.py which utilizes a model in combination with DaSiamRPN tracker

python volley_track.py --input_path assets\rally_men.mp4 --input_type video --onput_video_path Output/track.mp4 --model roboflow --marker circle --color yellow

5. Elif accuracy is what you are after, use detect.py. It calls the model on every frame.

python detect.py --input_path assets\rally_men.mp4 --input_type video --output_path Output/detect.mp4--model roboflow --marker circle --color yellow 

Output/track.mp4	Output/detect.mp4

Note that, it took volley_track.py 0.73 minutes to process the video, whereas detect.py completed in 2.75 minutes.

• You can get all list or arguments using -hcode> flag

python detect.py -h



For any additional quesitons feel free to take part in discussions or open an issue.