DESIGN AND IMPLEMENTATION OF FALL DETECTION USING CAMERA FOR AMBIENT ASSISTED LIVING

• Main Author: Hamka, Faudy
• Format: Final Project
• Language: Indonesia
• Online Access: https://digilib.itb.ac.id/gdl/view/40089
• Institution: Institut Teknologi Bandung

Indonesia population is about 265 million, nearly 25 million of which are elderly. The population of the elderly is expected to increase in the coming years. Therefore, it is necessary to fulfill the needs of health services for these elderly people. It can be done by implementing information and communication technology. One of the things that can be observed is the fall condition because the elderly is prone to fall. This final project aims to create a fall detection by using camera-based system in the elderly. Because the elderly is prone to fall, we have to provide SMS notifications to families so that when falling happens, it can be immediately followed up by them. The camera-based fall detection system is implemented using the OpenCV library and with the Python programming language. When the system successfully detects a fall, data in the form of time in the format of hours, minutes and seconds will be sent and stored into the MySQL database and can be accessed through the family application. At the same time, the system can provide a warning in the form of SMS notification. Sending this SMS notification is embedded in the Python program using Gammu SMS gateway. The SMS notification is sent to emergency numbers stored in the database. This camera-based fall detection system has an accuracy rate of 90% in a total of 60 experiments using datasets and direct experiments. Also, the average time needed to send an SMS notification from the time it is detected is 4.7615 seconds.