FPGA-Based Fall Detection System

As there is a high tendency of falling in the independent living of the elderly and the post-fall injury is very serious. It is necessary to get timely assistance when they fall. The main objective of this work is to build an FPGA-based hardware implementation of video-based fall detection system. First of all, the moving object model will be extracted through background subtraction based on Gaussian Mixture Models (GMM). Second, we judge whether there is a fall through the aspect ratio, the effective area ratio, and the change in the center of the human body. Finally, if the old person falls, the detection system will sound-light alarm and send message to the elderly family and community through GSM. The experimental results demonstrate the accuracy of this fall detection system is up to 95% indoor and this system satisfices the requirement of real-time.