OpenCv project to detect a person fall in videos with haarcascade
All packages used are in requirements.txt file
- python==2.7.7
- numpy==1.14.5
- opencv-python==3.4.1.15
- pylint==1.9.2
- scikit-learn==0.19.1
For each frame readed of the video corverted into gray, is removed the background, finded the contour and drawed the contours. If the heigh of the contour is lower than width, it may be a fall and we add 1 to a count, if the count is greater than 10, will be drawed a rectangle to the possible person fallen.
$ pip3 install -r requirements.txt
$ python3 fall.py
menbres du projet G1-G2 IAAP, fait à Centrale Lille de Novembre 20 - Fevrier 22.
Contact : projetiaap@gmail.com