Implementing YOLO technique in order to detect car accidents and to Detect an invasion of private territory

In this repository you could find several scripts in order to implement YOLOv3 algorithm on videos, images and camera.

At first, clone the repository:

git clone https://github.com/Terminateit/YOLOv3-Car-Accident-Detection.git

then go to the folder yolo-coco in the repository:

cd YOLOv3-Car-Accident-Detection/yolo-coco

Run the bash-script to download weights for models:

./download_yolov3_weights.sh

(May be, it the upper command will download the axel packet also (it's needed for boosting downloading from the net))

After that, you can run any of the scripts like:

python3 yolov3_<aim>.py

• yolov3_camera_SaFe_Territory.py - detect people and cars on the camera frames
• yolov3_CarAccidents.py - detect car accidents on the video frames (Dataset is given)
• yolov3_camera.py - detect any objects on the camera frames
• yolov3_detect.py - detect objects on the camera frames for RealSense Camera
• yolov3_image.py - detect objects on the image (put to input)
• yolov3_video.py - detect objects on the video frames

! If you have some troubles with OpenCV, try to do:

pip3 install opencv-python==4.2.0.32 && pip3 install opencv-contrib-python==4.2.0.32