/vechicle-counting-yolo

An algorithm that uses YoloV5 and DeepSORT to count and measure the number of vehicles in a video stream, it detects the vehicles with YoloV5 and tracks them with DeepSORT to maintain a count of unique vehicles in the video. It also calculates the frames per second (FPS) of the video stream.

Primary LanguagePythonMIT LicenseMIT

Vehicle Counting using Yolov5 and Deep Sort

Status License

On CPU - 12 to 15 FPS

Pre-requisites :

  1. Clone the Repository vehicle-counting-yolov5
git clone https://github.com/mahimairaja/vehicle-counting-yolov5.git

cd vehicle-counting-yolov5
  1. Clone the legacy Yolo-v5 Repository
git clone https://github.com/ultralytics/yolov5.git
  1. Install the libraries
pip install -r requirements.txt

Directory Structure :

After completing the above steps your directory should look like somewhat as of below structure

  • vehicle-counting-yolov5
    • deep_sort
    • yolov5
    • input.mp4
    • yolov5s.pt
    • tracker.py
    • requirements.txt

Run the algorithm

python tracker.py 
# This will download model weight - yolov5s.pt to base folder on first execution.

Feel free to conect with me...