運用YOLO技術於銀髮族之跌倒偵測及物品協尋研究

YOLO Technology for Silver Hair Fall Detection and Object Locator Research

張志忠(Jhih-Chung Chang)；廖彥博(Yen-Po Liao)

AIoT ； YOLO ； 智慧化 ； 跌倒偵測 ； AIoT ； YOLO ； intellectual ； fall detection

嶺東學報

50期（2023 / 12 / 01）

255 - 296

繁體中文;英文

隨著台灣人口逐漸老化，老年人記憶能力逐漸衰退，時常遺失東西在某個地方。而老人家大多希望在家養老，老人居家安全就成為社會關切的議題，所以獨居老人的照顧和協助也成為一般民眾或者政府機關都得一同面對與思考的問題。本計畫首先針對年長者容易遺失物品的問題，提出結合AIoT技術來進行遺失物品協尋裝置，可協助老人尋找原本遺失的物品位置。其次，老人跌倒是公共衛生及其臨床重要的議題之一，不僅造成老年人的身體傷害、住院甚至死亡，也帶給醫療體系及家庭龐大的照護負擔。計畫也將運用YOLO辨識技術實作跌倒行為偵測，藉由AIoT來輔助，可以省下許多人力，讓長照人員能做更有意義的關懷服務。亦可與偏鄉社區長照機構做結合，達到即時照護的功效。

As Taiwan's population ages, the memory of the elderly gradually deteriorates, and they often lose things somewhere. As most of the elderly want to age at home, home safety for the elderly has become an issue of social concern, so care and assistance for the elderly living alone has become an issue that both the general public and governmental agencies have to face and think about together. This program firstly focuses on the problem that the elderly tend to lose their belongings, and proposes the combination of AIoT technology to carry out the lost item search device, which can assist the elderly to find the location of the original lost items. Secondly, elderly falls are one of the most important issues in public health and clinical practice, which not only cause physical injuries, hospitalization, and even death of the elderly, but also bring a huge burden of care to the medical system and families. The plan will also use YOLO recognition technology to detect fall behavior, assisted by AIoT, which can save a lot of manpower and allow long-term care staff to do more meaningful care services. It can also be combined with long term care organizations in remote communities to achieve the effect of immediate care.