This Repository contains the Video-based Accident Detection System built in Python.
Packages Required | installation commands |
---|---|
cv2 | pip install opencv-python |
numpy | pip install numpy |
tensorflow | pip install tensorflow |
patool | pip install patool |
- Clone the repository into your system using the command "git clone https://github.com/snehal2841/Accident-detection-system.git "
- Install all the packages mentioned above.
- Navigate to the directory.
- python download.py
- By running above command, it will create an accdet.rar file, which you have to unzip in the same folder and copy its content to the path of main folder.
- python Object_detection_video.py
For training we have used following standard dataset.
- Oxford Drive Datasets
- ACD3( Accident Detection dataset, Unsupervised Traffic Accident Detection in First-Person Videos)
- CADP (Unsupervised Traffic Accident Detection in First-Person Videos)
The dataset for this project is a video. That is present in the repository as "testvideo.mp4".