My B.Sc. project at Amirkabir University of Technology, Advisor: Dr. Shiry
This is a simple system for implementation of an automatic door using RGB data and a Raspberry Pi 3 (or any other device running Python and OpenCV).
OpenCV Version: 3.4.1 Python Version: 2.7
The system first performs a Motion Detection using a simple (yet effective) background subtraction. Then naive estimations of the trajectories for moving objects performed and the door status is classified.
Using USB camera:
python motion_detection.py
Using recorded video:
python motion_detection.py --video ../data/1_trimmed_1.mp4
For Raspberry Pi:
python motion_detection_rp.py
add --min-area to set the threshold for minimum contour area of detected objects. 1200 is recommended for examples in data folder.
aoi_x1, aoi_y1, aoi_x2, aoi_y2 are parameters for setting (x,y) location for top left and bottom right of Area of Interest (which is the door).