This project utilizes MediaPipe and OpenCV to detect whether a person is standing or has fallen down based on their keypoints extracted from the MediaPipe Pose module. The program analyses the keypoint landmarks obtained from a webcam or surveillance camera feed to determine a person's posture.
The Python code in this repository employs the MediaPipe Pose module to:
- Capture frames from a camera source (webcam or surveillance camera).
- Process the frames to identify human poses and landmarks.
- Calculate the dual and single axis distance between specific body parts (optional functionality).
- Determine if a person is standing or has fallen down based on predefined criteria.
To run this project, ensure you have the following dependencies installed:
- Python 3.x
- OpenCV (
cv2
) - MediaPipe (
mediapipe
)
- Clone this repository to your local machine.
- Install the necessary dependencies.
- Run the
main.py
script.